Vehicle Detection

In this project we detect vehicles from the self driving car dashcam and draw bounding box around the vehicles. Vehicle detection is an import part for autonomous driving to avoid collisions and maintain proper distance on the road.

The steps involved for detection are:

  1. Feature extraction for window classifier - including HOG features, spatial binning, color historams
  2. Build classifier on the Vehicles and Non-Vehicles training data based on extracted features
  3. Sliding window search on each frame of the video for drawing the bounding box around vehicles
  4. Process the video frame by frame with the image processing pipeline, draw bounding box and re-assemble new video

Heatmaps are used for overlapping bounding boxes

In [45]:
import matplotlib.image as mpimg
import matplotlib.pyplot as plt

import numpy as np
import cv2
import glob

from skimage.feature import hog
from scipy.ndimage.measurements import label
from sklearn.preprocessing import StandardScaler
from sklearn.model_selection import train_test_split
from sklearn.svm import SVC
from scipy.ndimage.measurements import label
from sklearn.externals import joblib

%matplotlib inline

Preprocessed images and converted to jpg

Note: The Vehicles and Non-Vehicles dataset provided had images in .png format. These were converted to /jpg using handy linux convert utility.

ls -1 *.png | xargs -n 1 bash -c 'convert "$0" "${0%.png}.jpg"'
In [31]:
cars = glob.glob('./vehicles/*/*.jpg')
print(len(cars))

8792
In [32]:
non_cars = glob.glob('./non-vehicles/*/*.jpg')
print(len(non_cars))

8968
In [33]:
def plot_image_set(images, titles, image_type=None):
    f, (ax1, ax2) = plt.subplots(1, 2, figsize=(20,10))
    ax1.imshow(images[0], cmap=image_type)
    ax1.set_title(titles[0], fontsize=30)
    ax2.imshow(images[1], cmap=image_type)
    ax2.set_title(titles[1], fontsize=30)

Feature Extraction Parameters

After several round of trial and error, the following are the parameters chosen for HOG, Spatial Bins and Color Histograms:-

  • Color Channel - RGB
  • Orientation bins - 8
  • Pixels per cell - 8
  • Cells per block - 2
  • Histogram Bins - 48
  • Histogram Bins Range - (0, 256)
  • Spatial Image Size - (16, 16)
In [34]:
params = dict()
params['nbins'] = 48
params['bins_range'] = (0, 256)
params['resize_image_size'] = (16, 16)
params['orient'] = 8
params['pixels_per_cell'] = 8
params['cells_per_block'] = 2

Color histogram for each of the color channels

In [35]:
def color_hist(image, nbins=32, bins_range=(0, 256)):
    channel1_hist = np.histogram(image[:,:,0], bins=nbins, range=bins_range)
    channel2_hist = np.histogram(image[:,:,1], bins=nbins, range=bins_range)
    channel3_hist = np.histogram(image[:,:,2], bins=nbins, range=bins_range)
    hist_features = np.concatenate((channel1_hist[0], channel2_hist[0], channel3_hist[0]))
    return hist_features

Spatial binning of the images

In [36]:
def bin_spatial(image, size=(32, 32)):
    features = cv2.resize(image, size).ravel()
    return features

Hog Feature extraction

Hog features are extracted individually from each of the color channels. Features extracted from each color channels are appended and np.ravel() is used to return 1-D array of features of all the color channels.

In [37]:
def get_hog_features(image, orient, pix_per_cell, cell_per_block):
    hog_features = []
    for color_channel in range(image.shape[2]):
        channel_features = hog(image[:,:,color_channel], orientations=orient, pixels_per_cell=(pix_per_cell, pix_per_cell),
                           cells_per_block=(cell_per_block, cell_per_block), transform_sqrt=False, 
                           visualise=False, feature_vector=True)
        hog_features.append(channel_features)
    hog_features = np.ravel(hog_features)    
    return hog_features

Extract all the features

In [38]:
def extract_features(images, params):
    features = []
    for img_path in images:
        image = mpimg.imread(img_path)
        spatial_features = bin_spatial(image, size=params['resize_image_size'])
        hist_features = color_hist(image, nbins=params['nbins'], bins_range=params['bins_range'])
        hog_features = get_hog_features(image, params['orient'], params['pixels_per_cell'], params['cells_per_block'])
        features.append(np.concatenate((spatial_features, hist_features,hog_features)))
    return features

Utility function to extract features from a single image

In [39]:
def extra_single_features(image, params):
    spatial_features = bin_spatial(image, size=params['resize_image_size'])
    hist_features = color_hist(image, nbins=params['nbins'], bins_range=params['bins_range'])
    hog_features = get_hog_features(image, params['orient'], params['pixels_per_cell'], params['cells_per_block'])
    return np.concatenate((spatial_features, hist_features,hog_features))

Now we process Vehicles and Non-vehicles training data provided and extract all the features for classifying

In [40]:
car_features = extract_features(cars, params)
In [41]:
non_car_features = extract_features(non_cars, params)

Train/Test split

Do the train/test split for both the vehicle and non-vehicle data sets

In [42]:
def create_train_test_set(car_features, non_car_features):
    X = np.vstack((np.array(car_features), np.array(non_car_features))).astype(np.float64)                        
    X_scaler = StandardScaler().fit(X)
    scaled_X = X_scaler.transform(X)
    Y = np.hstack((np.ones(len(car_features)), np.zeros(len(non_car_features))))
    X_train, X_test, Y_train, Y_test = train_test_split(scaled_X, Y, test_size=0.2, random_state=1234)
    
    return X_train, Y_train, X_test, Y_test, X_scaler
In [43]:
X_train, Y_train, X_test, Y_test, X_scaler = create_train_test_set(car_features, non_car_features)

SVC Classifier

VC function of sklearn.svm was used for building the classifier. SVC was chosen instead of LinearSVC because it resulted in increased accuracy, due to the probabilistic output of SVC.

An overall accuracy of 99.6% was achieved

In [44]:
def train_test_svm(X_train, Y_train, X_test, Y_test):
    svc = SVC(C=1.0, probability=True)
    svc.fit(X_train, Y_train)
    
    print("Accuracy of classifier - {:.4f}".format(svc.score(X_test, Y_test)))
    return svc
In [46]:
clf = train_test_svm(X_train, Y_train, X_test, Y_test)

Accuracy of classifier - 0.9961

Save classifier

Classifier was picked for iterative tests

In [47]:
joblib.dump(clf, 'model.pkl')
Out[47]:
['model.pkl']

After lot of trial and error following parameters were selected for the bounding box and sliding window search

Three different window sizes were selected.

For small windows :-
      a) Window Size - (64, 64)
      b) X Start Stop - (640, 1280)
      c) Y Start Stop - (300, 450)
      d) Overlap - (0.6, 0.6)

For medium windows :-
      a) Window Size - (96, 96)
      b) X Start Stop - (640, 1280)
      c) Y Start Stop - (300, 500)
      d) Overlap - (0.8, 0.8)

For large windows :-
      a) Window Size - (128, 128)
      b) X Start Stop - (640, 1280)
      c) Y Start Stop - (400, 550)
      d) Overlap - (0.85, 0.85)

All the windows obtained from above setsp are passed to seach_window to get the actual hot boxes using the SVC classifier predictions. The classifier is used to predict the probability of a car being present in the image. If probability is greater than 0.5 then, car is present in the image.

Next, a heatmap is created using the hot boxes from above step. And Apply_threshold is called to remove some of the false positives present.

scipy.ndimage.measurements' label() function helps in labelling the thresholding regions. Finally draw_labeled_bboxes is used to draw bounding box around the detected car.

In [48]:
def draw_boxes(img, bboxes, color=(0, 0, 255), thick=6):
    imcopy = np.copy(img)
    for bbox in bboxes:
        cv2.rectangle(imcopy, bbox[0], bbox[1], color, thick)
    return imcopy
    
    
def slide_window(img, x_start_stop=[None, None], y_start_stop=[None, None], xy_window=(64, 64), xy_overlap=(0.5, 0.5)):
    if x_start_stop[0] == None:
        x_start_stop[0] = 0
    if x_start_stop[1] == None:
        x_start_stop[1] = img.shape[1]
    if y_start_stop[0] == None:
        y_start_stop[0] = 0
    if y_start_stop[1] == None:
        y_start_stop[1] = img.shape[0]

    xspan = x_start_stop[1] - x_start_stop[0]
    yspan = y_start_stop[1] - y_start_stop[0]
    
    nx_pix_per_step = np.int(xy_window[0]*(1 - xy_overlap[0]))
    ny_pix_per_step = np.int(xy_window[1]*(1 - xy_overlap[1]))
    
    nx_windows = np.int(xspan/nx_pix_per_step) - 1
    ny_windows = np.int(yspan/ny_pix_per_step) - 1
    
    window_list = []
    
    for ys in range(ny_windows):
        for xs in range(nx_windows):
            startx = xs*nx_pix_per_step + x_start_stop[0]
            endx = startx + xy_window[0]
            starty = ys*ny_pix_per_step + y_start_stop[0]
            endy = starty + xy_window[1]
            window_list.append(((startx, starty), (endx, endy)))
    
    return window_list
In [49]:
def search_windows(img, windows, clf, scaler, params):
    on_windows = []
    for window in windows:
        test_img = cv2.resize(img[window[0][1]:window[1][1], window[0][0]:window[1][0]], (64, 64))      
        features = extra_single_features(test_img, params)
        test_features = scaler.transform(np.array(features).reshape(1, -1))
        prediction = clf.predict(test_features)
        if prediction > 0.5:
            on_windows.append(window)
    return on_windows
In [50]:
all_test_images = glob.glob('test_images/*')
In [51]:
def add_heat(heatmap, bbox_list):
    for box in bbox_list:
        heatmap[box[0][1]:box[1][1], box[0][0]:box[1][0]] += 1

    return heatmap
    
def apply_threshold(heatmap, threshold):
    heatmap[heatmap <= threshold] = 0
    return heatmap
In [52]:
def draw_labeled_bboxes(img, labels):
    for car_number in range(1, labels[1]+1):
        nonzero = (labels[0] == car_number).nonzero()
        nonzeroy = np.array(nonzero[0])
        nonzerox = np.array(nonzero[1])

        bbox = ((np.min(nonzerox), np.min(nonzeroy)), (np.max(nonzerox), np.max(nonzeroy)))
        bbox_w, bbox_h = abs(bbox[1][0] - bbox[0][0]), abs(bbox[1][1] - bbox[0][1])
        if bbox[0][1] <= 300:
            continue
        if bbox[0][1] >= 600:
            continue
        if bbox[0][0] >= 1220:
            continue
        bbox_area = bbox_w * bbox_h
        if bbox_area <= 2500:
            continue
        cv2.rectangle(img, bbox[0], bbox[1], (0,0,255), 6)
    return img

Full image processing pipeline

Bounding boxes obtained in image pipeline in each frame were flickering a lot. To reduce the flickering Exponential Smoothening has been used. Heatmaps obtained from previous frame is given weightage of 0.2 and heatmaps from current frame are given weightage of 0.8 (values arrived after many experiments).

In [53]:
def image_pipeline(image, static_images=False):
    global prev_frame_heat
    image_copy = np.copy(image)
    
    windows1 = slide_window(image, x_start_stop=[640, 1280], y_start_stop=[300, 450], 
                xy_window=(64, 64), xy_overlap=(0.6, 0.6))
    
    windows2 = slide_window(image, x_start_stop=[640, 1280], y_start_stop=[300, 500], 
                xy_window=(96, 96), xy_overlap=(0.8, 0.8))
    
    windows3 = slide_window(image, x_start_stop=[640, 1280], y_start_stop=[400, 550], 
                xy_window=(128, 128), xy_overlap=(0.85, 0.85))

    windows = windows1 + windows2 + windows3 

    hot_windows = search_windows(image, windows, clf, X_scaler, params)                  

    heat_image = np.zeros_like(image[:,:,0]).astype(np.float)

    add_heat(heat_image, hot_windows)
    
    '''Apply Simple Exponential Filter'''
    if not static_images:
        if prev_frame_heat == None:
            prev_frame_heat = heat_image
        else:
            heat_image = (prev_frame_heat * 0.2) + (heat_image * 0.8)
            prev_frame_heat = heat_image
        
    heatmaps = apply_threshold(heat_image, 2)
    labels = label(heatmaps)  #from scipy.ndimage.measurements
    
    final_image = draw_labeled_bboxes(np.copy(image), labels)
    return final_image

False positive handling

To remove False positives following techniques were used:

  • draw_labeled_bboxes() function checks whether the minimum y-coordinate values is in between 300 and 600. If not, then don't draw the box. Also, minimum x-coordinate value should be less than 1220. One more check of the area of the box has been used to remove small False positve boxes. If the area of the box is less than 2500, then don't draw it.

  • Use of SVC with probabilitic prediction instead of normal LinearSVC also helped in removing many False positives.

  • Heatmaps and thresholding on those heatmaps also helped in removing False positives.

In [54]:
for image_path in all_test_images:
    test_img = mpimg.imread(image_path)
    res_img = image_pipeline(test_img, static_images=True)
    plot_image_set([test_img, res_img], titles=['Original Image', 'Vehicles'])

Inside look

Let's look inside the vehicle object detection and see the intermediate steps with multi-boxes and heatmaps

In [60]:
def image_pipeline_details(image, static_images=False):
    global prev_frame_heat
    image_copy = np.copy(image)
    
    windows1 = slide_window(image, x_start_stop=[640, 1280], y_start_stop=[300, 450], 
                xy_window=(64, 64), xy_overlap=(0.6, 0.6))
    
    windows2 = slide_window(image, x_start_stop=[640, 1280], y_start_stop=[300, 500], 
                xy_window=(96, 96), xy_overlap=(0.8, 0.8))
    
    windows3 = slide_window(image, x_start_stop=[640, 1280], y_start_stop=[400, 550], 
                xy_window=(128, 128), xy_overlap=(0.85, 0.85))

    windows = windows1 + windows2 + windows3 

    hot_windows = search_windows(image, windows, clf, X_scaler, params)                  

    heat_image = np.zeros_like(image[:,:,0]).astype(np.float)

    add_heat(heat_image, hot_windows)
    
    '''Apply Simple Exponential Filter'''
    if not static_images:
        if prev_frame_heat == None:
            prev_frame_heat = heat_image
        else:
            heat_image = (prev_frame_heat * 0.2) + (heat_image * 0.8)
            prev_frame_heat = heat_image
        
    heatmaps = apply_threshold(heat_image, 2)
    labels = label(heatmaps)
    
    final_image = draw_labeled_bboxes(np.copy(image), labels)
    hot_image = draw_boxes(np.copy(image), hot_windows)
    
    return final_image, heat_image, hot_image
In [61]:
def plot_image_set4(images, titles, image_type=None):
    f, (ax1, ax2, ax3, ax4) = plt.subplots(1, 4, figsize=(20,10))
    ax1.axis('off')
    ax1.imshow(images[0], cmap=image_type)
    ax1.set_title(titles[0], fontsize=20)
    ax2.axis('off')
    ax2.imshow(images[1], cmap=image_type)
    ax2.set_title(titles[1], fontsize=20)
    ax3.axis('off')
    ax3.imshow(images[2], cmap=image_type)
    ax3.set_title(titles[2], fontsize=20)
    ax4.axis('off')
    ax4.imshow(images[3], cmap=image_type)
    ax4.set_title(titles[3], fontsize=20)
In [62]:
prev_frame_heat = None
for image_path in all_test_images:
    test_img = mpimg.imread(image_path)
    res_img, heat_img, hot_img = image_pipeline_details(test_img, static_images=False)
    plot_image_set4([test_img, hot_img, heat_img, res_img], titles=['Original Image',"Multi-boxes", 'Heatmap', "Vehicles"])

Video processing

We will now process the video through the vehicle detection image pipeline and generate a new video with bounding boxes around the vehicles.

In [55]:
from moviepy.editor import VideoFileClip
from IPython.display import HTML
In [56]:
def process_image(image):
    result = image_pipeline(image)
    return result
In [57]:
prev_frame_heat = None
output_clip = 'project_video_result.mp4'
input_clip = VideoFileClip("project_video.mp4")
clip = input_clip.fl_image(process_image)
%time clip.write_videofile(output_clip, audio=False)

[MoviePy] >>>> Building video project_video_result.mp4
[MoviePy] Writing video project_video_result.mp4

  0%|          | 0/1261 [00:00<?, ?it/s]
  0%|          | 1/1261 [00:13<4:44:03, 13.53s/it]
  0%|          | 2/1261 [00:26<4:43:29, 13.51s/it]
  0%|          | 3/1261 [00:40<4:43:27, 13.52s/it]
  0%|          | 4/1261 [00:54<4:42:54, 13.50s/it]
  0%|          | 5/1261 [01:07<4:42:55, 13.52s/it]
  0%|          | 6/1261 [01:20<4:42:14, 13.49s/it]
  1%|          | 7/1261 [01:34<4:42:12, 13.50s/it]
  1%|          | 8/1261 [01:47<4:41:34, 13.48s/it]
  1%|          | 9/1261 [02:01<4:41:35, 13.49s/it]
  1%|          | 10/1261 [02:14<4:41:01, 13.48s/it]
  1%|          | 11/1261 [02:28<4:41:10, 13.50s/it]
  1%|          | 12/1261 [02:41<4:40:43, 13.49s/it]
  1%|          | 13/1261 [02:55<4:40:45, 13.50s/it]
  1%|          | 14/1261 [03:08<4:40:08, 13.48s/it]
  1%|          | 15/1261 [03:22<4:40:15, 13.50s/it]
  1%|▏         | 16/1261 [03:35<4:39:44, 13.48s/it]
  1%|▏         | 17/1261 [03:49<4:39:48, 13.50s/it]
  1%|▏         | 18/1261 [04:02<4:39:14, 13.48s/it]
  2%|▏         | 19/1261 [04:16<4:39:23, 13.50s/it]
  2%|▏         | 20/1261 [04:29<4:38:50, 13.48s/it]
  2%|▏         | 21/1261 [04:43<4:38:53, 13.49s/it]
  2%|▏         | 22/1261 [04:56<4:38:18, 13.48s/it]
  2%|▏         | 23/1261 [05:10<4:38:22, 13.49s/it]
  2%|▏         | 24/1261 [05:23<4:37:49, 13.48s/it]
  2%|▏         | 25/1261 [05:37<4:37:55, 13.49s/it]
  2%|▏         | 26/1261 [05:50<4:37:20, 13.47s/it]
  2%|▏         | 27/1261 [06:04<4:37:24, 13.49s/it]
  2%|▏         | 28/1261 [06:17<4:36:50, 13.47s/it]
  2%|▏         | 29/1261 [06:31<4:36:58, 13.49s/it]
  2%|▏         | 30/1261 [06:44<4:36:24, 13.47s/it]
  2%|▏         | 31/1261 [06:58<4:36:37, 13.49s/it]
  3%|▎         | 32/1261 [07:11<4:36:15, 13.49s/it]
  3%|▎         | 33/1261 [07:25<4:36:15, 13.50s/it]
  3%|▎         | 34/1261 [07:38<4:35:36, 13.48s/it]
  3%|▎         | 35/1261 [07:52<4:35:36, 13.49s/it]
  3%|▎         | 36/1261 [08:05<4:35:03, 13.47s/it]
  3%|▎         | 37/1261 [08:19<4:35:07, 13.49s/it]
  3%|▎         | 38/1261 [08:32<4:34:38, 13.47s/it]
  3%|▎         | 39/1261 [08:46<4:34:41, 13.49s/it]
  3%|▎         | 40/1261 [08:59<4:34:07, 13.47s/it]
  3%|▎         | 41/1261 [09:12<4:34:10, 13.48s/it]
  3%|▎         | 42/1261 [09:26<4:33:36, 13.47s/it]
  3%|▎         | 43/1261 [09:39<4:33:39, 13.48s/it]
  3%|▎         | 44/1261 [09:53<4:33:08, 13.47s/it]
  4%|▎         | 45/1261 [10:06<4:33:12, 13.48s/it]
  4%|▎         | 46/1261 [10:20<4:32:46, 13.47s/it]
  4%|▎         | 47/1261 [10:33<4:32:48, 13.48s/it]
  4%|▍         | 48/1261 [10:47<4:32:15, 13.47s/it]
  4%|▍         | 49/1261 [11:00<4:32:22, 13.48s/it]
  4%|▍         | 50/1261 [11:14<4:31:53, 13.47s/it]
  4%|▍         | 51/1261 [11:27<4:31:57, 13.49s/it]
  4%|▍         | 52/1261 [11:41<4:31:24, 13.47s/it]
  4%|▍         | 53/1261 [11:54<4:31:28, 13.48s/it]
  4%|▍         | 54/1261 [12:08<4:30:54, 13.47s/it]
  4%|▍         | 55/1261 [12:21<4:30:59, 13.48s/it]
  4%|▍         | 56/1261 [12:35<4:30:30, 13.47s/it]
  5%|▍         | 57/1261 [12:48<4:30:43, 13.49s/it]
  5%|▍         | 58/1261 [13:02<4:30:09, 13.47s/it]
  5%|▍         | 59/1261 [13:15<4:30:11, 13.49s/it]
  5%|▍         | 60/1261 [13:28<4:29:38, 13.47s/it]
  5%|▍         | 61/1261 [13:42<4:29:52, 13.49s/it]
  5%|▍         | 62/1261 [13:55<4:29:20, 13.48s/it]
  5%|▍         | 63/1261 [14:09<4:29:22, 13.49s/it]
  5%|▌         | 64/1261 [14:22<4:28:47, 13.47s/it]
  5%|▌         | 65/1261 [14:36<4:28:48, 13.49s/it]
  5%|▌         | 66/1261 [14:49<4:28:13, 13.47s/it]
  5%|▌         | 67/1261 [15:03<4:28:12, 13.48s/it]
  5%|▌         | 68/1261 [15:16<4:27:40, 13.46s/it]
  5%|▌         | 69/1261 [15:30<4:27:42, 13.47s/it]
  6%|▌         | 70/1261 [15:43<4:27:10, 13.46s/it]
  6%|▌         | 71/1261 [15:57<4:27:13, 13.47s/it]
  6%|▌         | 72/1261 [16:10<4:26:49, 13.46s/it]
  6%|▌         | 73/1261 [16:24<4:26:51, 13.48s/it]
  6%|▌         | 74/1261 [16:37<4:26:22, 13.46s/it]
  6%|▌         | 75/1261 [16:51<4:26:24, 13.48s/it]
  6%|▌         | 76/1261 [17:04<4:25:51, 13.46s/it]
  6%|▌         | 77/1261 [17:18<4:25:58, 13.48s/it]
  6%|▌         | 78/1261 [17:31<4:25:28, 13.46s/it]
  6%|▋         | 79/1261 [17:45<4:25:41, 13.49s/it]
  6%|▋         | 80/1261 [17:58<4:25:07, 13.47s/it]
  6%|▋         | 81/1261 [18:11<4:25:11, 13.48s/it]
  7%|▋         | 82/1261 [18:25<4:24:39, 13.47s/it]
  7%|▋         | 83/1261 [18:38<4:24:41, 13.48s/it]
  7%|▋         | 84/1261 [18:52<4:24:13, 13.47s/it]
  7%|▋         | 85/1261 [19:05<4:24:17, 13.48s/it]
  7%|▋         | 86/1261 [19:19<4:23:48, 13.47s/it]
  7%|▋         | 87/1261 [19:32<4:23:50, 13.48s/it]
  7%|▋         | 88/1261 [19:46<4:23:19, 13.47s/it]
  7%|▋         | 89/1261 [19:59<4:23:21, 13.48s/it]
  7%|▋         | 90/1261 [20:13<4:22:51, 13.47s/it]
  7%|▋         | 91/1261 [20:26<4:23:12, 13.50s/it]
  7%|▋         | 92/1261 [20:40<4:22:40, 13.48s/it]
  7%|▋         | 93/1261 [20:53<4:22:38, 13.49s/it]
  7%|▋         | 94/1261 [21:07<4:22:02, 13.47s/it]
  8%|▊         | 95/1261 [21:20<4:22:03, 13.49s/it]
  8%|▊         | 96/1261 [21:34<4:21:27, 13.47s/it]
  8%|▊         | 97/1261 [21:47<4:21:28, 13.48s/it]
  8%|▊         | 98/1261 [22:01<4:20:54, 13.46s/it]
  8%|▊         | 99/1261 [22:14<4:21:08, 13.48s/it]
  8%|▊         | 100/1261 [22:28<4:20:35, 13.47s/it]
  8%|▊         | 101/1261 [22:41<4:20:36, 13.48s/it]
  8%|▊         | 102/1261 [22:54<4:20:04, 13.46s/it]
  8%|▊         | 103/1261 [23:08<4:20:09, 13.48s/it]
  8%|▊         | 104/1261 [23:21<4:19:40, 13.47s/it]
  8%|▊         | 105/1261 [23:35<4:19:42, 13.48s/it]
  8%|▊         | 106/1261 [23:48<4:19:10, 13.46s/it]
  8%|▊         | 107/1261 [24:02<4:19:15, 13.48s/it]
  9%|▊         | 108/1261 [24:15<4:18:46, 13.47s/it]
  9%|▊         | 109/1261 [24:29<4:18:54, 13.49s/it]
  9%|▊         | 110/1261 [24:42<4:18:21, 13.47s/it]
  9%|▉         | 111/1261 [24:56<4:18:24, 13.48s/it]
  9%|▉         | 112/1261 [25:09<4:17:53, 13.47s/it]
  9%|▉         | 113/1261 [25:23<4:18:00, 13.48s/it]
  9%|▉         | 114/1261 [25:36<4:17:28, 13.47s/it]
  9%|▉         | 115/1261 [25:50<4:17:33, 13.48s/it]
  9%|▉         | 116/1261 [26:03<4:17:06, 13.47s/it]
  9%|▉         | 117/1261 [26:17<4:17:17, 13.49s/it]
  9%|▉         | 118/1261 [26:30<4:16:54, 13.49s/it]
  9%|▉         | 119/1261 [26:44<4:16:57, 13.50s/it]
 10%|▉         | 120/1261 [26:57<4:16:26, 13.49s/it]
 10%|▉         | 121/1261 [27:11<4:16:36, 13.51s/it]
 10%|▉         | 122/1261 [27:24<4:16:06, 13.49s/it]
 10%|▉         | 123/1261 [27:38<4:16:05, 13.50s/it]
 10%|▉         | 124/1261 [27:51<4:15:31, 13.48s/it]
 10%|▉         | 125/1261 [28:05<4:15:30, 13.50s/it]
 10%|▉         | 126/1261 [28:18<4:14:56, 13.48s/it]
 10%|█         | 127/1261 [28:32<4:14:57, 13.49s/it]
 10%|█         | 128/1261 [28:45<4:14:23, 13.47s/it]
 10%|█         | 129/1261 [28:59<4:14:26, 13.49s/it]
 10%|█         | 130/1261 [29:12<4:13:55, 13.47s/it]
 10%|█         | 131/1261 [29:25<4:14:02, 13.49s/it]
 10%|█         | 132/1261 [29:39<4:13:29, 13.47s/it]
 11%|█         | 133/1261 [29:52<4:13:38, 13.49s/it]
 11%|█         | 134/1261 [30:06<4:13:05, 13.47s/it]
 11%|█         | 135/1261 [30:19<4:13:10, 13.49s/it]
 11%|█         | 136/1261 [30:33<4:12:39, 13.48s/it]
 11%|█         | 137/1261 [30:46<4:12:41, 13.49s/it]
 11%|█         | 138/1261 [31:00<4:12:09, 13.47s/it]
 11%|█         | 139/1261 [31:13<4:12:12, 13.49s/it]
 11%|█         | 140/1261 [31:27<4:11:42, 13.47s/it]
 11%|█         | 141/1261 [31:40<4:11:45, 13.49s/it]
 11%|█▏        | 142/1261 [31:54<4:11:15, 13.47s/it]
 11%|█▏        | 143/1261 [32:07<4:11:23, 13.49s/it]
 11%|█▏        | 144/1261 [32:21<4:10:51, 13.48s/it]
 11%|█▏        | 145/1261 [32:34<4:10:56, 13.49s/it]
 12%|█▏        | 146/1261 [32:48<4:10:23, 13.47s/it]
 12%|█▏        | 147/1261 [33:01<4:10:26, 13.49s/it]
 12%|█▏        | 148/1261 [33:15<4:09:55, 13.47s/it]
 12%|█▏        | 149/1261 [33:28<4:09:56, 13.49s/it]
 12%|█▏        | 150/1261 [33:42<4:09:27, 13.47s/it]
 12%|█▏        | 151/1261 [33:55<4:09:37, 13.49s/it]
 12%|█▏        | 152/1261 [34:09<4:09:07, 13.48s/it]
 12%|█▏        | 153/1261 [34:22<4:09:22, 13.50s/it]
 12%|█▏        | 154/1261 [34:36<4:08:55, 13.49s/it]
 12%|█▏        | 155/1261 [34:49<4:08:56, 13.51s/it]
 12%|█▏        | 156/1261 [35:03<4:08:27, 13.49s/it]
 12%|█▏        | 157/1261 [35:16<4:08:28, 13.50s/it]
 13%|█▎        | 158/1261 [35:30<4:07:58, 13.49s/it]
 13%|█▎        | 159/1261 [35:43<4:07:58, 13.50s/it]
 13%|█▎        | 160/1261 [35:57<4:07:23, 13.48s/it]
 13%|█▎        | 161/1261 [36:10<4:07:25, 13.50s/it]
 13%|█▎        | 162/1261 [36:24<4:06:56, 13.48s/it]
 13%|█▎        | 163/1261 [36:37<4:06:57, 13.50s/it]
 13%|█▎        | 164/1261 [36:50<4:06:29, 13.48s/it]
 13%|█▎        | 165/1261 [37:04<4:06:33, 13.50s/it]
 13%|█▎        | 166/1261 [37:17<4:06:05, 13.48s/it]
 13%|█▎        | 167/1261 [37:31<4:06:09, 13.50s/it]
 13%|█▎        | 168/1261 [37:44<4:05:39, 13.49s/it]
 13%|█▎        | 169/1261 [37:58<4:05:43, 13.50s/it]
 13%|█▎        | 170/1261 [38:11<4:05:22, 13.49s/it]
 14%|█▎        | 171/1261 [38:25<4:05:24, 13.51s/it]
 14%|█▎        | 172/1261 [38:39<4:04:59, 13.50s/it]
 14%|█▎        | 173/1261 [38:52<4:04:59, 13.51s/it]
 14%|█▍        | 174/1261 [39:05<4:04:27, 13.49s/it]
 14%|█▍        | 175/1261 [39:19<4:04:30, 13.51s/it]
 14%|█▍        | 176/1261 [39:32<4:03:57, 13.49s/it]
 14%|█▍        | 177/1261 [39:46<4:03:58, 13.50s/it]
 14%|█▍        | 178/1261 [39:59<4:03:28, 13.49s/it]
 14%|█▍        | 179/1261 [40:13<4:03:32, 13.50s/it]
 14%|█▍        | 180/1261 [40:26<4:03:04, 13.49s/it]
 14%|█▍        | 181/1261 [40:40<4:03:14, 13.51s/it]
 14%|█▍        | 182/1261 [40:54<4:02:47, 13.50s/it]
 15%|█▍        | 183/1261 [41:07<4:02:45, 13.51s/it]
 15%|█▍        | 184/1261 [41:21<4:02:14, 13.50s/it]
 15%|█▍        | 185/1261 [41:34<4:02:13, 13.51s/it]
 15%|█▍        | 186/1261 [41:47<4:01:39, 13.49s/it]
 15%|█▍        | 187/1261 [42:01<4:01:40, 13.50s/it]
 15%|█▍        | 188/1261 [42:14<4:01:12, 13.49s/it]
 15%|█▍        | 189/1261 [42:28<4:01:16, 13.50s/it]
 15%|█▌        | 190/1261 [42:41<4:00:49, 13.49s/it]
 15%|█▌        | 191/1261 [42:55<4:00:49, 13.50s/it]
 15%|█▌        | 192/1261 [43:08<4:00:20, 13.49s/it]
 15%|█▌        | 193/1261 [43:22<4:00:24, 13.51s/it]
 15%|█▌        | 194/1261 [43:35<3:59:54, 13.49s/it]
 15%|█▌        | 195/1261 [43:49<3:59:56, 13.51s/it]
 16%|█▌        | 196/1261 [44:02<3:59:26, 13.49s/it]
 16%|█▌        | 197/1261 [44:16<3:59:34, 13.51s/it]
 16%|█▌        | 198/1261 [44:29<3:59:09, 13.50s/it]
 16%|█▌        | 199/1261 [44:43<3:59:05, 13.51s/it]
 16%|█▌        | 200/1261 [44:56<3:58:34, 13.49s/it]
 16%|█▌        | 201/1261 [45:10<3:58:38, 13.51s/it]
 16%|█▌        | 202/1261 [45:23<3:58:07, 13.49s/it]
 16%|█▌        | 203/1261 [45:37<3:58:08, 13.51s/it]
 16%|█▌        | 204/1261 [45:50<3:57:39, 13.49s/it]
 16%|█▋        | 205/1261 [46:04<3:57:42, 13.51s/it]
 16%|█▋        | 206/1261 [46:17<3:57:16, 13.49s/it]
 16%|█▋        | 207/1261 [46:31<3:57:18, 13.51s/it]
 16%|█▋        | 208/1261 [46:45<3:56:56, 13.50s/it]
 17%|█▋        | 209/1261 [46:58<3:56:57, 13.51s/it]
 17%|█▋        | 210/1261 [47:12<3:56:28, 13.50s/it]
 17%|█▋        | 211/1261 [47:25<3:56:36, 13.52s/it]
 17%|█▋        | 212/1261 [47:39<3:56:04, 13.50s/it]
 17%|█▋        | 213/1261 [47:52<3:56:04, 13.52s/it]
 17%|█▋        | 214/1261 [48:06<3:55:32, 13.50s/it]
 17%|█▋        | 215/1261 [48:19<3:55:34, 13.51s/it]
 17%|█▋        | 216/1261 [48:33<3:55:03, 13.50s/it]
 17%|█▋        | 217/1261 [48:46<3:55:05, 13.51s/it]
 17%|█▋        | 218/1261 [49:00<3:54:35, 13.50s/it]
 17%|█▋        | 219/1261 [49:13<3:54:43, 13.52s/it]
 17%|█▋        | 220/1261 [49:27<3:54:11, 13.50s/it]
 18%|█▊        | 221/1261 [49:40<3:54:12, 13.51s/it]
 18%|█▊        | 222/1261 [49:54<3:53:42, 13.50s/it]
 18%|█▊        | 223/1261 [50:07<3:53:42, 13.51s/it]
 18%|█▊        | 224/1261 [50:21<3:53:14, 13.50s/it]
 18%|█▊        | 225/1261 [50:34<3:53:16, 13.51s/it]
 18%|█▊        | 226/1261 [50:48<3:52:47, 13.50s/it]
 18%|█▊        | 227/1261 [51:01<3:52:54, 13.51s/it]
 18%|█▊        | 228/1261 [51:15<3:52:33, 13.51s/it]
 18%|█▊        | 229/1261 [51:28<3:52:32, 13.52s/it]
 18%|█▊        | 230/1261 [51:42<3:52:01, 13.50s/it]
 18%|█▊        | 231/1261 [51:55<3:52:01, 13.52s/it]
 18%|█▊        | 232/1261 [52:09<3:51:30, 13.50s/it]
 18%|█▊        | 233/1261 [52:22<3:51:34, 13.52s/it]
 19%|█▊        | 234/1261 [52:36<3:51:04, 13.50s/it]
 19%|█▊        | 235/1261 [52:49<3:51:05, 13.51s/it]
 19%|█▊        | 236/1261 [53:03<3:50:35, 13.50s/it]
 19%|█▉        | 237/1261 [53:16<3:50:38, 13.51s/it]
 19%|█▉        | 238/1261 [53:30<3:50:13, 13.50s/it]
 19%|█▉        | 239/1261 [53:43<3:50:17, 13.52s/it]
 19%|█▉        | 240/1261 [53:57<3:49:47, 13.50s/it]
 19%|█▉        | 241/1261 [54:10<3:49:57, 13.53s/it]
 19%|█▉        | 242/1261 [54:24<3:49:27, 13.51s/it]
 19%|█▉        | 243/1261 [54:37<3:49:28, 13.52s/it]
 19%|█▉        | 244/1261 [54:51<3:48:57, 13.51s/it]
 19%|█▉        | 245/1261 [55:04<3:48:58, 13.52s/it]
 20%|█▉        | 246/1261 [55:18<3:48:31, 13.51s/it]
 20%|█▉        | 247/1261 [55:31<3:48:38, 13.53s/it]
 20%|█▉        | 248/1261 [55:45<3:48:07, 13.51s/it]
 20%|█▉        | 249/1261 [55:58<3:48:07, 13.53s/it]
 20%|█▉        | 250/1261 [56:12<3:47:40, 13.51s/it]
 20%|█▉        | 251/1261 [56:26<3:47:45, 13.53s/it]
 20%|█▉        | 252/1261 [56:39<3:47:15, 13.51s/it]
 20%|██        | 253/1261 [56:53<3:47:16, 13.53s/it]
 20%|██        | 254/1261 [57:06<3:46:45, 13.51s/it]
 20%|██        | 255/1261 [57:20<3:46:47, 13.53s/it]
 20%|██        | 256/1261 [57:33<3:46:17, 13.51s/it]
 20%|██        | 257/1261 [57:47<3:46:20, 13.53s/it]
 20%|██        | 258/1261 [58:00<3:45:52, 13.51s/it]
 21%|██        | 259/1261 [58:14<3:45:56, 13.53s/it]
 21%|██        | 260/1261 [58:27<3:45:29, 13.52s/it]
 21%|██        | 261/1261 [58:41<3:45:29, 13.53s/it]
 21%|██        | 262/1261 [58:54<3:44:59, 13.51s/it]
 21%|██        | 263/1261 [59:08<3:45:01, 13.53s/it]
 21%|██        | 264/1261 [59:21<3:44:35, 13.52s/it]
 21%|██        | 265/1261 [59:35<3:44:38, 13.53s/it]
 21%|██        | 266/1261 [59:48<3:44:08, 13.52s/it]
 21%|██        | 267/1261 [1:00:02<3:44:10, 13.53s/it]
 21%|██▏       | 268/1261 [1:00:15<3:43:42, 13.52s/it]
 21%|██▏       | 269/1261 [1:00:29<3:43:44, 13.53s/it]
 21%|██▏       | 270/1261 [1:00:42<3:43:15, 13.52s/it]
 21%|██▏       | 271/1261 [1:00:56<3:43:22, 13.54s/it]
 22%|██▏       | 272/1261 [1:01:09<3:42:52, 13.52s/it]
 22%|██▏       | 273/1261 [1:01:23<3:42:52, 13.54s/it]
 22%|██▏       | 274/1261 [1:01:37<3:42:24, 13.52s/it]
 22%|██▏       | 275/1261 [1:01:50<3:42:24, 13.53s/it]
 22%|██▏       | 276/1261 [1:02:04<3:41:53, 13.52s/it]
 22%|██▏       | 277/1261 [1:02:17<3:41:58, 13.54s/it]
 22%|██▏       | 278/1261 [1:02:31<3:41:27, 13.52s/it]
 22%|██▏       | 279/1261 [1:02:44<3:41:26, 13.53s/it]
 22%|██▏       | 280/1261 [1:02:58<3:40:57, 13.51s/it]
 22%|██▏       | 281/1261 [1:03:11<3:41:03, 13.53s/it]
 22%|██▏       | 282/1261 [1:03:25<3:40:37, 13.52s/it]
 22%|██▏       | 283/1261 [1:03:38<3:40:34, 13.53s/it]
 23%|██▎       | 284/1261 [1:03:52<3:40:06, 13.52s/it]
 23%|██▎       | 285/1261 [1:04:05<3:40:07, 13.53s/it]
 23%|██▎       | 286/1261 [1:04:19<3:39:40, 13.52s/it]
 23%|██▎       | 287/1261 [1:04:32<3:39:39, 13.53s/it]
 23%|██▎       | 288/1261 [1:04:46<3:39:13, 13.52s/it]
 23%|██▎       | 289/1261 [1:04:59<3:39:13, 13.53s/it]
 23%|██▎       | 290/1261 [1:05:13<3:38:46, 13.52s/it]
 23%|██▎       | 291/1261 [1:05:26<3:38:53, 13.54s/it]
 23%|██▎       | 292/1261 [1:05:40<3:38:29, 13.53s/it]
 23%|██▎       | 293/1261 [1:05:54<3:38:31, 13.55s/it]
 23%|██▎       | 294/1261 [1:06:07<3:38:00, 13.53s/it]
 23%|██▎       | 295/1261 [1:06:21<3:38:01, 13.54s/it]
 23%|██▎       | 296/1261 [1:06:34<3:37:31, 13.52s/it]
 24%|██▎       | 297/1261 [1:06:48<3:37:31, 13.54s/it]
 24%|██▎       | 298/1261 [1:07:01<3:37:01, 13.52s/it]
 24%|██▎       | 299/1261 [1:07:15<3:37:04, 13.54s/it]
 24%|██▍       | 300/1261 [1:07:28<3:36:37, 13.53s/it]
 24%|██▍       | 301/1261 [1:07:42<3:36:43, 13.55s/it]
 24%|██▍       | 302/1261 [1:07:55<3:36:15, 13.53s/it]
 24%|██▍       | 303/1261 [1:08:09<3:36:16, 13.54s/it]
 24%|██▍       | 304/1261 [1:08:22<3:35:47, 13.53s/it]
 24%|██▍       | 305/1261 [1:08:36<3:35:45, 13.54s/it]
 24%|██▍       | 306/1261 [1:08:49<3:35:15, 13.52s/it]
 24%|██▍       | 307/1261 [1:09:03<3:35:16, 13.54s/it]
 24%|██▍       | 308/1261 [1:09:17<3:34:48, 13.52s/it]
 25%|██▍       | 309/1261 [1:09:30<3:34:49, 13.54s/it]
 25%|██▍       | 310/1261 [1:09:44<3:34:22, 13.53s/it]
 25%|██▍       | 311/1261 [1:09:57<3:34:21, 13.54s/it]
 25%|██▍       | 312/1261 [1:10:11<3:34:01, 13.53s/it]
 25%|██▍       | 313/1261 [1:10:24<3:34:01, 13.55s/it]
 25%|██▍       | 314/1261 [1:10:38<3:33:29, 13.53s/it]
 25%|██▍       | 315/1261 [1:10:51<3:33:29, 13.54s/it]
 25%|██▌       | 316/1261 [1:11:05<3:33:01, 13.52s/it]
 25%|██▌       | 317/1261 [1:11:18<3:33:04, 13.54s/it]
 25%|██▌       | 318/1261 [1:11:32<3:32:35, 13.53s/it]
 25%|██▌       | 319/1261 [1:11:45<3:32:38, 13.54s/it]
 25%|██▌       | 320/1261 [1:11:59<3:32:12, 13.53s/it]
 25%|██▌       | 321/1261 [1:12:13<3:32:14, 13.55s/it]
 26%|██▌       | 322/1261 [1:12:26<3:31:43, 13.53s/it]
 26%|██▌       | 323/1261 [1:12:40<3:31:43, 13.54s/it]
 26%|██▌       | 324/1261 [1:12:53<3:31:16, 13.53s/it]
 26%|██▌       | 325/1261 [1:13:07<3:31:15, 13.54s/it]
 26%|██▌       | 326/1261 [1:13:20<3:30:46, 13.53s/it]
 26%|██▌       | 327/1261 [1:13:34<3:30:47, 13.54s/it]
 26%|██▌       | 328/1261 [1:13:47<3:30:21, 13.53s/it]
 26%|██▌       | 329/1261 [1:14:01<3:30:23, 13.54s/it]
 26%|██▌       | 330/1261 [1:14:14<3:29:57, 13.53s/it]
 26%|██▌       | 331/1261 [1:14:28<3:30:03, 13.55s/it]
 26%|██▋       | 332/1261 [1:14:41<3:29:37, 13.54s/it]
 26%|██▋       | 333/1261 [1:14:55<3:29:36, 13.55s/it]
 26%|██▋       | 334/1261 [1:15:09<3:29:05, 13.53s/it]
 27%|██▋       | 335/1261 [1:15:22<3:29:06, 13.55s/it]
 27%|██▋       | 336/1261 [1:15:36<3:28:40, 13.54s/it]
 27%|██▋       | 337/1261 [1:15:49<3:28:39, 13.55s/it]
 27%|██▋       | 338/1261 [1:16:03<3:28:08, 13.53s/it]
 27%|██▋       | 339/1261 [1:16:16<3:28:09, 13.55s/it]
 27%|██▋       | 340/1261 [1:16:30<3:27:43, 13.53s/it]
 27%|██▋       | 341/1261 [1:16:43<3:27:43, 13.55s/it]
 27%|██▋       | 342/1261 [1:16:57<3:27:14, 13.53s/it]
 27%|██▋       | 343/1261 [1:17:10<3:27:15, 13.55s/it]
 27%|██▋       | 344/1261 [1:17:24<3:26:48, 13.53s/it]
 27%|██▋       | 345/1261 [1:17:37<3:26:48, 13.55s/it]
 27%|██▋       | 346/1261 [1:17:51<3:26:18, 13.53s/it]
 28%|██▊       | 347/1261 [1:18:05<3:26:19, 13.54s/it]
 28%|██▊       | 348/1261 [1:18:18<3:25:59, 13.54s/it]
 28%|██▊       | 349/1261 [1:18:32<3:25:58, 13.55s/it]
 28%|██▊       | 350/1261 [1:18:45<3:25:29, 13.53s/it]
 28%|██▊       | 351/1261 [1:18:59<3:25:26, 13.55s/it]
 28%|██▊       | 352/1261 [1:19:12<3:24:58, 13.53s/it]
 28%|██▊       | 353/1261 [1:19:26<3:24:57, 13.54s/it]
 28%|██▊       | 354/1261 [1:19:39<3:24:51, 13.55s/it]
 28%|██▊       | 355/1261 [1:19:53<3:24:46, 13.56s/it]
 28%|██▊       | 356/1261 [1:20:06<3:24:15, 13.54s/it]
 28%|██▊       | 357/1261 [1:20:20<3:24:14, 13.56s/it]
 28%|██▊       | 358/1261 [1:20:34<3:23:43, 13.54s/it]
 28%|██▊       | 359/1261 [1:20:47<3:23:41, 13.55s/it]
 29%|██▊       | 360/1261 [1:21:01<3:23:15, 13.54s/it]
 29%|██▊       | 361/1261 [1:21:14<3:23:16, 13.55s/it]
 29%|██▊       | 362/1261 [1:21:28<3:22:48, 13.54s/it]
 29%|██▉       | 363/1261 [1:21:41<3:22:15, 13.51s/it]
 29%|██▉       | 364/1261 [1:21:55<3:21:42, 13.49s/it]
 29%|██▉       | 365/1261 [1:22:08<3:21:29, 13.49s/it]
 29%|██▉       | 366/1261 [1:22:22<3:21:09, 13.49s/it]
 29%|██▉       | 367/1261 [1:22:35<3:21:00, 13.49s/it]
 29%|██▉       | 368/1261 [1:22:49<3:20:42, 13.49s/it]
 29%|██▉       | 369/1261 [1:23:02<3:20:32, 13.49s/it]
 29%|██▉       | 370/1261 [1:23:16<3:20:14, 13.48s/it]
 29%|██▉       | 371/1261 [1:23:29<3:20:05, 13.49s/it]
 30%|██▉       | 372/1261 [1:23:42<3:19:41, 13.48s/it]
 30%|██▉       | 373/1261 [1:23:56<3:19:25, 13.47s/it]
 30%|██▉       | 374/1261 [1:24:09<3:19:04, 13.47s/it]
 30%|██▉       | 375/1261 [1:24:23<3:18:51, 13.47s/it]
 30%|██▉       | 376/1261 [1:24:36<3:18:34, 13.46s/it]
 30%|██▉       | 377/1261 [1:24:50<3:18:21, 13.46s/it]
 30%|██▉       | 378/1261 [1:25:03<3:18:03, 13.46s/it]
 30%|███       | 379/1261 [1:25:17<3:17:52, 13.46s/it]
 30%|███       | 380/1261 [1:25:30<3:17:34, 13.46s/it]
 30%|███       | 381/1261 [1:25:44<3:17:24, 13.46s/it]
 30%|███       | 382/1261 [1:25:57<3:17:06, 13.45s/it]
 30%|███       | 383/1261 [1:26:10<3:16:55, 13.46s/it]
 30%|███       | 384/1261 [1:26:24<3:16:37, 13.45s/it]
 31%|███       | 385/1261 [1:26:37<3:16:25, 13.45s/it]
 31%|███       | 386/1261 [1:26:51<3:16:08, 13.45s/it]
 31%|███       | 387/1261 [1:27:04<3:16:05, 13.46s/it]
 31%|███       | 388/1261 [1:27:18<3:15:44, 13.45s/it]
 31%|███       | 389/1261 [1:27:31<3:15:32, 13.45s/it]
 31%|███       | 390/1261 [1:27:45<3:15:12, 13.45s/it]
 31%|███       | 391/1261 [1:27:58<3:15:06, 13.46s/it]
 31%|███       | 392/1261 [1:28:12<3:14:48, 13.45s/it]
 31%|███       | 393/1261 [1:28:25<3:14:37, 13.45s/it]
 31%|███       | 394/1261 [1:28:38<3:14:19, 13.45s/it]
 31%|███▏      | 395/1261 [1:28:52<3:14:14, 13.46s/it]
 31%|███▏      | 396/1261 [1:29:05<3:14:00, 13.46s/it]
 31%|███▏      | 397/1261 [1:29:19<3:13:49, 13.46s/it]
 32%|███▏      | 398/1261 [1:29:32<3:13:33, 13.46s/it]
 32%|███▏      | 399/1261 [1:29:46<3:13:21, 13.46s/it]
 32%|███▏      | 400/1261 [1:29:59<3:13:01, 13.45s/it]
 32%|███▏      | 401/1261 [1:30:13<3:12:49, 13.45s/it]
 32%|███▏      | 402/1261 [1:30:26<3:12:31, 13.45s/it]
 32%|███▏      | 403/1261 [1:30:40<3:12:30, 13.46s/it]
 32%|███▏      | 404/1261 [1:30:53<3:12:11, 13.46s/it]
 32%|███▏      | 405/1261 [1:31:06<3:11:58, 13.46s/it]
 32%|███▏      | 406/1261 [1:31:20<3:11:41, 13.45s/it]
 32%|███▏      | 407/1261 [1:31:33<3:11:31, 13.46s/it]
 32%|███▏      | 408/1261 [1:31:47<3:11:20, 13.46s/it]
 32%|███▏      | 409/1261 [1:32:00<3:11:05, 13.46s/it]
 33%|███▎      | 410/1261 [1:32:14<3:10:47, 13.45s/it]
 33%|███▎      | 411/1261 [1:32:27<3:10:37, 13.46s/it]
 33%|███▎      | 412/1261 [1:32:41<3:10:19, 13.45s/it]
 33%|███▎      | 413/1261 [1:32:54<3:10:14, 13.46s/it]
 33%|███▎      | 414/1261 [1:33:08<3:10:01, 13.46s/it]
 33%|███▎      | 415/1261 [1:33:21<3:09:55, 13.47s/it]
 33%|███▎      | 416/1261 [1:33:35<3:09:38, 13.47s/it]
 33%|███▎      | 417/1261 [1:33:48<3:09:29, 13.47s/it]
 33%|███▎      | 418/1261 [1:34:01<3:09:12, 13.47s/it]
 33%|███▎      | 419/1261 [1:34:15<3:08:57, 13.46s/it]
 33%|███▎      | 420/1261 [1:34:28<3:08:35, 13.46s/it]
 33%|███▎      | 421/1261 [1:34:42<3:08:27, 13.46s/it]
 33%|███▎      | 422/1261 [1:34:55<3:08:09, 13.46s/it]
 34%|███▎      | 423/1261 [1:35:09<3:07:57, 13.46s/it]
 34%|███▎      | 424/1261 [1:35:22<3:07:48, 13.46s/it]
 34%|███▎      | 425/1261 [1:35:36<3:07:32, 13.46s/it]
 34%|███▍      | 426/1261 [1:35:49<3:07:12, 13.45s/it]
 34%|███▍      | 427/1261 [1:36:03<3:06:57, 13.45s/it]
 34%|███▍      | 428/1261 [1:36:16<3:06:46, 13.45s/it]
 34%|███▍      | 429/1261 [1:36:29<3:06:35, 13.46s/it]
 34%|███▍      | 430/1261 [1:36:43<3:06:17, 13.45s/it]
 34%|███▍      | 431/1261 [1:36:56<3:06:02, 13.45s/it]
 34%|███▍      | 432/1261 [1:37:10<3:05:46, 13.45s/it]
 34%|███▍      | 433/1261 [1:37:23<3:05:32, 13.45s/it]
 34%|███▍      | 434/1261 [1:37:37<3:05:14, 13.44s/it]
 34%|███▍      | 435/1261 [1:37:50<3:05:01, 13.44s/it]
 35%|███▍      | 436/1261 [1:38:04<3:04:43, 13.43s/it]
 35%|███▍      | 437/1261 [1:38:17<3:04:40, 13.45s/it]
 35%|███▍      | 438/1261 [1:38:30<3:04:22, 13.44s/it]
 35%|███▍      | 439/1261 [1:38:44<3:04:09, 13.44s/it]
 35%|███▍      | 440/1261 [1:38:57<3:03:53, 13.44s/it]
 35%|███▍      | 441/1261 [1:39:11<3:03:42, 13.44s/it]
 35%|███▌      | 442/1261 [1:39:24<3:03:24, 13.44s/it]
 35%|███▌      | 443/1261 [1:39:38<3:03:18, 13.45s/it]
 35%|███▌      | 444/1261 [1:39:51<3:03:06, 13.45s/it]
 35%|███▌      | 445/1261 [1:40:05<3:02:52, 13.45s/it]
 35%|███▌      | 446/1261 [1:40:18<3:02:32, 13.44s/it]
 35%|███▌      | 447/1261 [1:40:31<3:02:27, 13.45s/it]
 36%|███▌      | 448/1261 [1:40:45<3:02:06, 13.44s/it]
 36%|███▌      | 449/1261 [1:40:58<3:01:53, 13.44s/it]
 36%|███▌      | 450/1261 [1:41:12<3:01:46, 13.45s/it]
 36%|███▌      | 451/1261 [1:41:25<3:01:37, 13.45s/it]
 36%|███▌      | 452/1261 [1:41:39<3:01:17, 13.45s/it]
 36%|███▌      | 453/1261 [1:41:52<3:01:04, 13.45s/it]
 36%|███▌      | 454/1261 [1:42:06<3:00:45, 13.44s/it]
 36%|███▌      | 455/1261 [1:42:19<3:00:33, 13.44s/it]
 36%|███▌      | 456/1261 [1:42:32<3:00:15, 13.43s/it]
 36%|███▌      | 457/1261 [1:42:46<3:00:02, 13.44s/it]
 36%|███▋      | 458/1261 [1:42:59<2:59:44, 13.43s/it]
 36%|███▋      | 459/1261 [1:43:13<2:59:33, 13.43s/it]
 36%|███▋      | 460/1261 [1:43:26<2:59:17, 13.43s/it]
 37%|███▋      | 461/1261 [1:43:40<2:59:07, 13.43s/it]
 37%|███▋      | 462/1261 [1:43:53<2:58:51, 13.43s/it]
 37%|███▋      | 463/1261 [1:44:06<2:58:43, 13.44s/it]
 37%|███▋      | 464/1261 [1:44:20<2:58:26, 13.43s/it]
 37%|███▋      | 465/1261 [1:44:33<2:58:15, 13.44s/it]
 37%|███▋      | 466/1261 [1:44:47<2:57:59, 13.43s/it]
 37%|███▋      | 467/1261 [1:45:00<2:57:55, 13.44s/it]
 37%|███▋      | 468/1261 [1:45:14<2:57:42, 13.45s/it]
 37%|███▋      | 469/1261 [1:45:27<2:57:35, 13.45s/it]
 37%|███▋      | 470/1261 [1:45:41<2:57:19, 13.45s/it]
 37%|███▋      | 471/1261 [1:45:54<2:57:09, 13.46s/it]
 37%|███▋      | 472/1261 [1:46:07<2:56:53, 13.45s/it]
 38%|███▊      | 473/1261 [1:46:21<2:56:38, 13.45s/it]
 38%|███▊      | 474/1261 [1:46:34<2:56:18, 13.44s/it]
 38%|███▊      | 475/1261 [1:46:48<2:56:04, 13.44s/it]
 38%|███▊      | 476/1261 [1:47:01<2:55:46, 13.43s/it]
 38%|███▊      | 477/1261 [1:47:15<2:55:34, 13.44s/it]
 38%|███▊      | 478/1261 [1:47:28<2:55:16, 13.43s/it]
 38%|███▊      | 479/1261 [1:47:42<2:55:05, 13.43s/it]
 38%|███▊      | 480/1261 [1:47:55<2:54:49, 13.43s/it]
 38%|███▊      | 481/1261 [1:48:08<2:54:51, 13.45s/it]
 38%|███▊      | 482/1261 [1:48:22<2:54:33, 13.45s/it]
 38%|███▊      | 483/1261 [1:48:35<2:54:20, 13.44s/it]
 38%|███▊      | 484/1261 [1:48:49<2:54:02, 13.44s/it]
 38%|███▊      | 485/1261 [1:49:02<2:53:50, 13.44s/it]
 39%|███▊      | 486/1261 [1:49:16<2:53:34, 13.44s/it]
 39%|███▊      | 487/1261 [1:49:29<2:53:27, 13.45s/it]
 39%|███▊      | 488/1261 [1:49:43<2:53:15, 13.45s/it]
 39%|███▉      | 489/1261 [1:49:56<2:53:02, 13.45s/it]
 39%|███▉      | 490/1261 [1:50:09<2:52:45, 13.44s/it]
 39%|███▉      | 491/1261 [1:50:23<2:52:35, 13.45s/it]
 39%|███▉      | 492/1261 [1:50:36<2:52:17, 13.44s/it]
 39%|███▉      | 493/1261 [1:50:50<2:52:18, 13.46s/it]
 39%|███▉      | 494/1261 [1:51:03<2:52:04, 13.46s/it]
 39%|███▉      | 495/1261 [1:51:17<2:51:57, 13.47s/it]
 39%|███▉      | 496/1261 [1:51:30<2:51:38, 13.46s/it]
 39%|███▉      | 497/1261 [1:51:44<2:51:29, 13.47s/it]
 39%|███▉      | 498/1261 [1:51:57<2:51:13, 13.46s/it]
 40%|███▉      | 499/1261 [1:52:11<2:51:04, 13.47s/it]
 40%|███▉      | 500/1261 [1:52:24<2:50:40, 13.46s/it]
 40%|███▉      | 501/1261 [1:52:37<2:50:23, 13.45s/it]
 40%|███▉      | 502/1261 [1:52:51<2:50:02, 13.44s/it]
 40%|███▉      | 503/1261 [1:53:04<2:49:49, 13.44s/it]
 40%|███▉      | 504/1261 [1:53:18<2:49:32, 13.44s/it]
 40%|████      | 505/1261 [1:53:31<2:49:22, 13.44s/it]
 40%|████      | 506/1261 [1:53:45<2:49:04, 13.44s/it]
 40%|████      | 507/1261 [1:53:58<2:48:51, 13.44s/it]
 40%|████      | 508/1261 [1:54:12<2:48:35, 13.43s/it]
 40%|████      | 509/1261 [1:54:25<2:48:23, 13.44s/it]
 40%|████      | 510/1261 [1:54:38<2:48:06, 13.43s/it]
 41%|████      | 511/1261 [1:54:52<2:48:00, 13.44s/it]
 41%|████      | 512/1261 [1:55:05<2:47:44, 13.44s/it]
 41%|████      | 513/1261 [1:55:19<2:47:41, 13.45s/it]
 41%|████      | 514/1261 [1:55:32<2:47:21, 13.44s/it]
 41%|████      | 515/1261 [1:55:46<2:47:06, 13.44s/it]
 41%|████      | 516/1261 [1:55:59<2:46:49, 13.44s/it]
 41%|████      | 517/1261 [1:56:13<2:46:41, 13.44s/it]
 41%|████      | 518/1261 [1:56:26<2:46:21, 13.43s/it]
 41%|████      | 519/1261 [1:56:39<2:46:06, 13.43s/it]
 41%|████      | 520/1261 [1:56:53<2:45:57, 13.44s/it]
 41%|████▏     | 521/1261 [1:57:06<2:45:48, 13.44s/it]
 41%|████▏     | 522/1261 [1:57:20<2:45:32, 13.44s/it]
 41%|████▏     | 523/1261 [1:57:33<2:45:21, 13.44s/it]
 42%|████▏     | 524/1261 [1:57:47<2:45:06, 13.44s/it]
 42%|████▏     | 525/1261 [1:58:00<2:44:56, 13.45s/it]
 42%|████▏     | 526/1261 [1:58:13<2:44:43, 13.45s/it]
 42%|████▏     | 527/1261 [1:58:27<2:44:30, 13.45s/it]
 42%|████▏     | 528/1261 [1:58:40<2:44:08, 13.44s/it]
 42%|████▏     | 529/1261 [1:58:54<2:43:53, 13.43s/it]
 42%|████▏     | 530/1261 [1:59:07<2:43:34, 13.43s/it]
 42%|████▏     | 531/1261 [1:59:21<2:43:21, 13.43s/it]
 42%|████▏     | 532/1261 [1:59:34<2:43:06, 13.42s/it]
 42%|████▏     | 533/1261 [1:59:47<2:42:54, 13.43s/it]
 42%|████▏     | 534/1261 [2:00:01<2:42:38, 13.42s/it]
 42%|████▏     | 535/1261 [2:00:14<2:42:28, 13.43s/it]
 43%|████▎     | 536/1261 [2:00:28<2:42:12, 13.42s/it]
 43%|████▎     | 537/1261 [2:00:41<2:42:00, 13.43s/it]
 43%|████▎     | 538/1261 [2:00:55<2:41:44, 13.42s/it]
 43%|████▎     | 539/1261 [2:01:08<2:41:37, 13.43s/it]
 43%|████▎     | 540/1261 [2:01:21<2:41:26, 13.43s/it]
 43%|████▎     | 541/1261 [2:01:35<2:41:21, 13.45s/it]
 43%|████▎     | 542/1261 [2:01:48<2:41:06, 13.44s/it]
 43%|████▎     | 543/1261 [2:02:02<2:40:56, 13.45s/it]
 43%|████▎     | 544/1261 [2:02:15<2:40:41, 13.45s/it]
 43%|████▎     | 545/1261 [2:02:29<2:40:31, 13.45s/it]
 43%|████▎     | 546/1261 [2:02:42<2:40:10, 13.44s/it]
 43%|████▎     | 547/1261 [2:02:56<2:39:55, 13.44s/it]
 43%|████▎     | 548/1261 [2:03:09<2:39:42, 13.44s/it]
 44%|████▎     | 549/1261 [2:03:22<2:39:30, 13.44s/it]
 44%|████▎     | 550/1261 [2:03:36<2:39:13, 13.44s/it]
 44%|████▎     | 551/1261 [2:03:49<2:39:07, 13.45s/it]
 44%|████▍     | 552/1261 [2:04:03<2:38:56, 13.45s/it]
 44%|████▍     | 553/1261 [2:04:16<2:38:49, 13.46s/it]
 44%|████▍     | 554/1261 [2:04:30<2:38:35, 13.46s/it]
 44%|████▍     | 555/1261 [2:04:43<2:38:26, 13.47s/it]
 44%|████▍     | 556/1261 [2:04:57<2:38:13, 13.47s/it]
 44%|████▍     | 557/1261 [2:05:10<2:37:55, 13.46s/it]
 44%|████▍     | 558/1261 [2:05:24<2:37:35, 13.45s/it]
 44%|████▍     | 559/1261 [2:05:37<2:37:19, 13.45s/it]
 44%|████▍     | 560/1261 [2:05:50<2:37:00, 13.44s/it]
 44%|████▍     | 561/1261 [2:06:04<2:36:47, 13.44s/it]
 45%|████▍     | 562/1261 [2:06:17<2:36:31, 13.44s/it]
 45%|████▍     | 563/1261 [2:06:31<2:36:19, 13.44s/it]
 45%|████▍     | 564/1261 [2:06:44<2:36:08, 13.44s/it]
 45%|████▍     | 565/1261 [2:06:58<2:36:02, 13.45s/it]
 45%|████▍     | 566/1261 [2:07:11<2:35:48, 13.45s/it]
 45%|████▍     | 567/1261 [2:07:25<2:35:36, 13.45s/it]
 45%|████▌     | 568/1261 [2:07:38<2:35:24, 13.45s/it]
 45%|████▌     | 569/1261 [2:07:52<2:35:16, 13.46s/it]
 45%|████▌     | 570/1261 [2:08:05<2:34:56, 13.45s/it]
 45%|████▌     | 571/1261 [2:08:18<2:34:41, 13.45s/it]
 45%|████▌     | 572/1261 [2:08:32<2:34:22, 13.44s/it]
 45%|████▌     | 573/1261 [2:08:45<2:34:07, 13.44s/it]
 46%|████▌     | 574/1261 [2:08:59<2:33:48, 13.43s/it]
 46%|████▌     | 575/1261 [2:09:12<2:33:35, 13.43s/it]
 46%|████▌     | 576/1261 [2:09:26<2:33:21, 13.43s/it]
 46%|████▌     | 577/1261 [2:09:39<2:33:18, 13.45s/it]
 46%|████▌     | 578/1261 [2:09:52<2:33:06, 13.45s/it]
 46%|████▌     | 579/1261 [2:10:06<2:32:56, 13.46s/it]
 46%|████▌     | 580/1261 [2:10:19<2:32:43, 13.46s/it]
 46%|████▌     | 581/1261 [2:10:33<2:32:32, 13.46s/it]
 46%|████▌     | 582/1261 [2:10:46<2:32:21, 13.46s/it]
 46%|████▌     | 583/1261 [2:11:00<2:32:14, 13.47s/it]
 46%|████▋     | 584/1261 [2:11:13<2:32:03, 13.48s/it]
 46%|████▋     | 585/1261 [2:11:27<2:32:00, 13.49s/it]
 46%|████▋     | 586/1261 [2:11:40<2:31:44, 13.49s/it]
 47%|████▋     | 587/1261 [2:11:54<2:31:31, 13.49s/it]
 47%|████▋     | 588/1261 [2:12:07<2:31:11, 13.48s/it]
 47%|████▋     | 589/1261 [2:12:21<2:30:56, 13.48s/it]
 47%|████▋     | 590/1261 [2:12:34<2:30:36, 13.47s/it]
 47%|████▋     | 591/1261 [2:12:48<2:30:23, 13.47s/it]
 47%|████▋     | 592/1261 [2:13:01<2:30:02, 13.46s/it]
 47%|████▋     | 593/1261 [2:13:15<2:29:52, 13.46s/it]
 47%|████▋     | 594/1261 [2:13:28<2:29:32, 13.45s/it]
 47%|████▋     | 595/1261 [2:13:41<2:29:18, 13.45s/it]
 47%|████▋     | 596/1261 [2:13:55<2:29:02, 13.45s/it]
 47%|████▋     | 597/1261 [2:14:08<2:28:49, 13.45s/it]
 47%|████▋     | 598/1261 [2:14:22<2:28:31, 13.44s/it]
 48%|████▊     | 599/1261 [2:14:35<2:28:28, 13.46s/it]
 48%|████▊     | 600/1261 [2:14:49<2:28:08, 13.45s/it]
 48%|████▊     | 601/1261 [2:15:02<2:27:58, 13.45s/it]
 48%|████▊     | 602/1261 [2:15:16<2:27:39, 13.44s/it]
 48%|████▊     | 603/1261 [2:15:29<2:27:25, 13.44s/it]
 48%|████▊     | 604/1261 [2:15:42<2:27:08, 13.44s/it]
 48%|████▊     | 605/1261 [2:15:56<2:26:55, 13.44s/it]
 48%|████▊     | 606/1261 [2:16:09<2:26:50, 13.45s/it]
 48%|████▊     | 607/1261 [2:16:23<2:26:55, 13.48s/it]
 48%|████▊     | 608/1261 [2:16:36<2:26:41, 13.48s/it]
 48%|████▊     | 609/1261 [2:16:50<2:26:40, 13.50s/it]
 48%|████▊     | 610/1261 [2:17:03<2:26:19, 13.49s/it]
 48%|████▊     | 611/1261 [2:17:17<2:26:17, 13.50s/it]
 49%|████▊     | 612/1261 [2:17:30<2:26:00, 13.50s/it]
 49%|████▊     | 613/1261 [2:17:44<2:25:55, 13.51s/it]
 49%|████▊     | 614/1261 [2:17:57<2:25:29, 13.49s/it]
 49%|████▉     | 615/1261 [2:18:11<2:25:25, 13.51s/it]
 49%|████▉     | 616/1261 [2:18:24<2:25:01, 13.49s/it]
 49%|████▉     | 617/1261 [2:18:38<2:24:57, 13.50s/it]
 49%|████▉     | 618/1261 [2:18:51<2:24:31, 13.49s/it]
 49%|████▉     | 619/1261 [2:19:05<2:24:27, 13.50s/it]
 49%|████▉     | 620/1261 [2:19:18<2:24:04, 13.49s/it]
 49%|████▉     | 621/1261 [2:19:32<2:24:00, 13.50s/it]
 49%|████▉     | 622/1261 [2:19:45<2:23:36, 13.48s/it]
 49%|████▉     | 623/1261 [2:19:59<2:23:34, 13.50s/it]
 49%|████▉     | 624/1261 [2:20:12<2:23:13, 13.49s/it]
 50%|████▉     | 625/1261 [2:20:26<2:23:08, 13.50s/it]
 50%|████▉     | 626/1261 [2:20:39<2:22:41, 13.48s/it]
 50%|████▉     | 627/1261 [2:20:53<2:22:36, 13.50s/it]
 50%|████▉     | 628/1261 [2:21:06<2:22:11, 13.48s/it]
 50%|████▉     | 629/1261 [2:21:20<2:22:07, 13.49s/it]
 50%|████▉     | 630/1261 [2:21:33<2:21:40, 13.47s/it]
 50%|█████     | 631/1261 [2:21:47<2:21:41, 13.49s/it]
 50%|█████     | 632/1261 [2:22:00<2:21:19, 13.48s/it]
 50%|█████     | 633/1261 [2:22:14<2:21:17, 13.50s/it]
 50%|█████     | 634/1261 [2:22:27<2:21:00, 13.49s/it]
 50%|█████     | 635/1261 [2:22:41<2:20:55, 13.51s/it]
 50%|█████     | 636/1261 [2:22:54<2:20:33, 13.49s/it]
 51%|█████     | 637/1261 [2:23:08<2:20:29, 13.51s/it]
 51%|█████     | 638/1261 [2:23:21<2:20:07, 13.50s/it]
 51%|█████     | 639/1261 [2:23:35<2:20:03, 13.51s/it]
 51%|█████     | 640/1261 [2:23:48<2:19:40, 13.50s/it]
 51%|█████     | 641/1261 [2:24:02<2:19:35, 13.51s/it]
 51%|█████     | 642/1261 [2:24:15<2:19:11, 13.49s/it]
 51%|█████     | 643/1261 [2:24:29<2:19:07, 13.51s/it]
 51%|█████     | 644/1261 [2:24:42<2:18:44, 13.49s/it]
 51%|█████     | 645/1261 [2:24:56<2:18:40, 13.51s/it]
 51%|█████     | 646/1261 [2:25:09<2:18:17, 13.49s/it]
 51%|█████▏    | 647/1261 [2:25:23<2:18:13, 13.51s/it]
 51%|█████▏    | 648/1261 [2:25:36<2:17:51, 13.49s/it]
 51%|█████▏    | 649/1261 [2:25:50<2:17:46, 13.51s/it]
 52%|█████▏    | 650/1261 [2:26:03<2:17:24, 13.49s/it]
 52%|█████▏    | 651/1261 [2:26:17<2:17:21, 13.51s/it]
 52%|█████▏    | 652/1261 [2:26:30<2:17:00, 13.50s/it]
 52%|█████▏    | 653/1261 [2:26:44<2:16:58, 13.52s/it]
 52%|█████▏    | 654/1261 [2:26:57<2:16:38, 13.51s/it]
 52%|█████▏    | 655/1261 [2:27:11<2:16:33, 13.52s/it]
 52%|█████▏    | 656/1261 [2:27:24<2:16:08, 13.50s/it]
 52%|█████▏    | 657/1261 [2:27:38<2:16:05, 13.52s/it]
 52%|█████▏    | 658/1261 [2:27:51<2:15:41, 13.50s/it]
 52%|█████▏    | 659/1261 [2:28:05<2:15:38, 13.52s/it]
 52%|█████▏    | 660/1261 [2:28:18<2:15:15, 13.50s/it]
 52%|█████▏    | 661/1261 [2:28:32<2:15:13, 13.52s/it]
 52%|█████▏    | 662/1261 [2:28:45<2:14:49, 13.50s/it]
 53%|█████▎    | 663/1261 [2:28:59<2:14:45, 13.52s/it]
 53%|█████▎    | 664/1261 [2:29:12<2:14:23, 13.51s/it]
 53%|█████▎    | 665/1261 [2:29:26<2:14:19, 13.52s/it]
 53%|█████▎    | 666/1261 [2:29:39<2:13:55, 13.50s/it]
 53%|█████▎    | 667/1261 [2:29:53<2:13:50, 13.52s/it]
 53%|█████▎    | 668/1261 [2:30:06<2:13:25, 13.50s/it]
 53%|█████▎    | 669/1261 [2:30:20<2:13:20, 13.51s/it]
 53%|█████▎    | 670/1261 [2:30:33<2:12:55, 13.49s/it]
 53%|█████▎    | 671/1261 [2:30:47<2:12:50, 13.51s/it]
 53%|█████▎    | 672/1261 [2:31:00<2:12:27, 13.49s/it]
 53%|█████▎    | 673/1261 [2:31:14<2:12:23, 13.51s/it]
 53%|█████▎    | 674/1261 [2:31:27<2:12:00, 13.49s/it]
 54%|█████▎    | 675/1261 [2:31:41<2:11:56, 13.51s/it]
 54%|█████▎    | 676/1261 [2:31:54<2:11:34, 13.50s/it]
 54%|█████▎    | 677/1261 [2:32:08<2:11:29, 13.51s/it]
 54%|█████▍    | 678/1261 [2:32:21<2:11:06, 13.49s/it]
 54%|█████▍    | 679/1261 [2:32:35<2:11:01, 13.51s/it]
 54%|█████▍    | 680/1261 [2:32:49<2:10:41, 13.50s/it]
 54%|█████▍    | 681/1261 [2:33:02<2:10:38, 13.51s/it]
 54%|█████▍    | 682/1261 [2:33:16<2:10:14, 13.50s/it]
 54%|█████▍    | 683/1261 [2:33:29<2:10:09, 13.51s/it]
 54%|█████▍    | 684/1261 [2:33:43<2:09:45, 13.49s/it]
 54%|█████▍    | 685/1261 [2:33:56<2:09:41, 13.51s/it]
 54%|█████▍    | 686/1261 [2:34:10<2:09:20, 13.50s/it]
 54%|█████▍    | 687/1261 [2:34:23<2:09:15, 13.51s/it]
 55%|█████▍    | 688/1261 [2:34:37<2:08:53, 13.50s/it]
 55%|█████▍    | 689/1261 [2:34:50<2:08:48, 13.51s/it]
 55%|█████▍    | 690/1261 [2:35:04<2:08:26, 13.50s/it]
 55%|█████▍    | 691/1261 [2:35:17<2:08:26, 13.52s/it]
 55%|█████▍    | 692/1261 [2:35:31<2:08:05, 13.51s/it]
 55%|█████▍    | 693/1261 [2:35:44<2:08:02, 13.53s/it]
 55%|█████▌    | 694/1261 [2:35:58<2:07:39, 13.51s/it]
 55%|█████▌    | 695/1261 [2:36:11<2:07:34, 13.52s/it]
 55%|█████▌    | 696/1261 [2:36:25<2:07:13, 13.51s/it]
 55%|█████▌    | 697/1261 [2:36:38<2:07:08, 13.53s/it]
 55%|█████▌    | 698/1261 [2:36:52<2:06:44, 13.51s/it]
 55%|█████▌    | 699/1261 [2:37:05<2:06:38, 13.52s/it]
 56%|█████▌    | 700/1261 [2:37:19<2:06:15, 13.50s/it]
 56%|█████▌    | 701/1261 [2:37:32<2:06:11, 13.52s/it]
 56%|█████▌    | 702/1261 [2:37:46<2:05:48, 13.50s/it]
 56%|█████▌    | 703/1261 [2:37:59<2:05:42, 13.52s/it]
 56%|█████▌    | 704/1261 [2:38:13<2:05:22, 13.50s/it]
 56%|█████▌    | 705/1261 [2:38:26<2:05:20, 13.53s/it]
 56%|█████▌    | 706/1261 [2:38:40<2:04:59, 13.51s/it]
 56%|█████▌    | 707/1261 [2:38:53<2:04:55, 13.53s/it]
 56%|█████▌    | 708/1261 [2:39:07<2:04:34, 13.52s/it]
 56%|█████▌    | 709/1261 [2:39:20<2:04:30, 13.53s/it]
 56%|█████▋    | 710/1261 [2:39:34<2:04:08, 13.52s/it]
 56%|█████▋    | 711/1261 [2:39:48<2:04:05, 13.54s/it]
 56%|█████▋    | 712/1261 [2:40:01<2:03:40, 13.52s/it]
 57%|█████▋    | 713/1261 [2:40:15<2:03:36, 13.53s/it]
 57%|█████▋    | 714/1261 [2:40:28<2:03:13, 13.52s/it]
 57%|█████▋    | 715/1261 [2:40:42<2:03:07, 13.53s/it]
 57%|█████▋    | 716/1261 [2:40:55<2:02:44, 13.51s/it]
 57%|█████▋    | 717/1261 [2:41:09<2:02:39, 13.53s/it]
 57%|█████▋    | 718/1261 [2:41:22<2:02:17, 13.51s/it]
 57%|█████▋    | 719/1261 [2:41:36<2:02:12, 13.53s/it]
 57%|█████▋    | 720/1261 [2:41:49<2:01:49, 13.51s/it]
 57%|█████▋    | 721/1261 [2:42:03<2:01:47, 13.53s/it]
 57%|█████▋    | 722/1261 [2:42:16<2:01:26, 13.52s/it]
 57%|█████▋    | 723/1261 [2:42:30<2:01:21, 13.53s/it]
 57%|█████▋    | 724/1261 [2:42:43<2:00:59, 13.52s/it]
 57%|█████▋    | 725/1261 [2:42:57<2:00:53, 13.53s/it]
 58%|█████▊    | 726/1261 [2:43:10<2:00:30, 13.51s/it]
 58%|█████▊    | 727/1261 [2:43:24<2:00:24, 13.53s/it]
 58%|█████▊    | 728/1261 [2:43:37<2:00:00, 13.51s/it]
 58%|█████▊    | 729/1261 [2:43:51<1:59:56, 13.53s/it]
 58%|█████▊    | 730/1261 [2:44:04<1:59:35, 13.51s/it]
 58%|█████▊    | 731/1261 [2:44:18<1:59:31, 13.53s/it]
 58%|█████▊    | 732/1261 [2:44:31<1:59:09, 13.52s/it]
 58%|█████▊    | 733/1261 [2:44:45<1:59:04, 13.53s/it]
 58%|█████▊    | 734/1261 [2:44:58<1:58:41, 13.51s/it]
 58%|█████▊    | 735/1261 [2:45:12<1:58:35, 13.53s/it]
 58%|█████▊    | 736/1261 [2:45:26<1:58:16, 13.52s/it]
 58%|█████▊    | 737/1261 [2:45:39<1:58:13, 13.54s/it]
 59%|█████▊    | 738/1261 [2:45:53<1:57:51, 13.52s/it]
 59%|█████▊    | 739/1261 [2:46:06<1:57:45, 13.54s/it]
 59%|█████▊    | 740/1261 [2:46:20<1:57:23, 13.52s/it]
 59%|█████▉    | 741/1261 [2:46:33<1:57:17, 13.53s/it]
 59%|█████▉    | 742/1261 [2:46:47<1:56:54, 13.51s/it]
 59%|█████▉    | 743/1261 [2:47:00<1:56:48, 13.53s/it]
 59%|█████▉    | 744/1261 [2:47:14<1:56:27, 13.51s/it]
 59%|█████▉    | 745/1261 [2:47:27<1:56:23, 13.53s/it]
 59%|█████▉    | 746/1261 [2:47:41<1:56:01, 13.52s/it]
 59%|█████▉    | 747/1261 [2:47:54<1:55:57, 13.54s/it]
 59%|█████▉    | 748/1261 [2:48:08<1:55:34, 13.52s/it]
 59%|█████▉    | 749/1261 [2:48:21<1:55:29, 13.53s/it]
 59%|█████▉    | 750/1261 [2:48:35<1:55:07, 13.52s/it]
 60%|█████▉    | 751/1261 [2:48:48<1:55:05, 13.54s/it]
 60%|█████▉    | 752/1261 [2:49:02<1:54:44, 13.52s/it]
 60%|█████▉    | 753/1261 [2:49:16<1:54:39, 13.54s/it]
 60%|█████▉    | 754/1261 [2:49:29<1:54:19, 13.53s/it]
 60%|█████▉    | 755/1261 [2:49:43<1:54:12, 13.54s/it]
 60%|█████▉    | 756/1261 [2:49:56<1:53:48, 13.52s/it]
 60%|██████    | 757/1261 [2:50:10<1:53:41, 13.54s/it]
 60%|██████    | 758/1261 [2:50:23<1:53:17, 13.51s/it]
 60%|██████    | 759/1261 [2:50:37<1:53:11, 13.53s/it]
 60%|██████    | 760/1261 [2:50:50<1:52:49, 13.51s/it]
 60%|██████    | 761/1261 [2:51:04<1:52:43, 13.53s/it]
 60%|██████    | 762/1261 [2:51:17<1:52:21, 13.51s/it]
 61%|██████    | 763/1261 [2:51:31<1:52:15, 13.53s/it]
 61%|██████    | 764/1261 [2:51:44<1:51:54, 13.51s/it]
 61%|██████    | 765/1261 [2:51:58<1:51:49, 13.53s/it]
 61%|██████    | 766/1261 [2:52:11<1:51:29, 13.51s/it]
 61%|██████    | 767/1261 [2:52:25<1:51:23, 13.53s/it]
 61%|██████    | 768/1261 [2:52:38<1:51:01, 13.51s/it]
 61%|██████    | 769/1261 [2:52:52<1:50:55, 13.53s/it]
 61%|██████    | 770/1261 [2:53:05<1:50:33, 13.51s/it]
 61%|██████    | 771/1261 [2:53:19<1:50:28, 13.53s/it]
 61%|██████    | 772/1261 [2:53:32<1:50:06, 13.51s/it]
 61%|██████▏   | 773/1261 [2:53:46<1:50:00, 13.53s/it]
 61%|██████▏   | 774/1261 [2:53:59<1:49:39, 13.51s/it]
 61%|██████▏   | 775/1261 [2:54:13<1:49:34, 13.53s/it]
 62%|██████▏   | 776/1261 [2:54:26<1:49:12, 13.51s/it]
 62%|██████▏   | 777/1261 [2:54:40<1:49:06, 13.53s/it]
 62%|██████▏   | 778/1261 [2:54:54<1:48:45, 13.51s/it]
 62%|██████▏   | 779/1261 [2:55:07<1:48:41, 13.53s/it]
 62%|██████▏   | 780/1261 [2:55:21<1:48:21, 13.52s/it]
 62%|██████▏   | 781/1261 [2:55:34<1:48:19, 13.54s/it]
 62%|██████▏   | 782/1261 [2:55:48<1:47:57, 13.52s/it]
 62%|██████▏   | 783/1261 [2:56:01<1:47:50, 13.54s/it]
 62%|██████▏   | 784/1261 [2:56:15<1:47:29, 13.52s/it]
 62%|██████▏   | 785/1261 [2:56:28<1:47:22, 13.54s/it]
 62%|██████▏   | 786/1261 [2:56:42<1:47:00, 13.52s/it]
 62%|██████▏   | 787/1261 [2:56:55<1:46:53, 13.53s/it]
 62%|██████▏   | 788/1261 [2:57:09<1:46:31, 13.51s/it]
 63%|██████▎   | 789/1261 [2:57:22<1:46:25, 13.53s/it]
 63%|██████▎   | 790/1261 [2:57:36<1:46:03, 13.51s/it]
 63%|██████▎   | 791/1261 [2:57:49<1:45:57, 13.53s/it]
 63%|██████▎   | 792/1261 [2:58:03<1:45:36, 13.51s/it]
 63%|██████▎   | 793/1261 [2:58:16<1:45:29, 13.53s/it]
 63%|██████▎   | 794/1261 [2:58:30<1:45:09, 13.51s/it]
 63%|██████▎   | 795/1261 [2:58:43<1:45:04, 13.53s/it]
 63%|██████▎   | 796/1261 [2:58:57<1:44:42, 13.51s/it]
 63%|██████▎   | 797/1261 [2:59:10<1:44:35, 13.53s/it]
 63%|██████▎   | 798/1261 [2:59:24<1:44:14, 13.51s/it]
 63%|██████▎   | 799/1261 [2:59:38<1:44:08, 13.52s/it]
 63%|██████▎   | 800/1261 [2:59:51<1:43:49, 13.51s/it]
 64%|██████▎   | 801/1261 [3:00:05<1:43:41, 13.53s/it]
 64%|██████▎   | 802/1261 [3:00:18<1:43:19, 13.51s/it]
 64%|██████▎   | 803/1261 [3:00:32<1:43:13, 13.52s/it]
 64%|██████▍   | 804/1261 [3:00:45<1:42:51, 13.51s/it]
 64%|██████▍   | 805/1261 [3:00:59<1:42:46, 13.52s/it]
 64%|██████▍   | 806/1261 [3:01:12<1:42:25, 13.51s/it]
 64%|██████▍   | 807/1261 [3:01:26<1:42:19, 13.52s/it]
 64%|██████▍   | 808/1261 [3:01:39<1:41:57, 13.50s/it]
 64%|██████▍   | 809/1261 [3:01:53<1:41:51, 13.52s/it]
 64%|██████▍   | 810/1261 [3:02:06<1:41:30, 13.50s/it]
 64%|██████▍   | 811/1261 [3:02:20<1:41:29, 13.53s/it]
 64%|██████▍   | 812/1261 [3:02:33<1:41:08, 13.52s/it]
 64%|██████▍   | 813/1261 [3:02:47<1:41:01, 13.53s/it]
 65%|██████▍   | 814/1261 [3:03:00<1:40:39, 13.51s/it]
 65%|██████▍   | 815/1261 [3:03:14<1:40:32, 13.53s/it]
 65%|██████▍   | 816/1261 [3:03:27<1:40:10, 13.51s/it]
 65%|██████▍   | 817/1261 [3:03:41<1:40:04, 13.52s/it]
 65%|██████▍   | 818/1261 [3:03:54<1:39:42, 13.50s/it]
 65%|██████▍   | 819/1261 [3:04:08<1:39:36, 13.52s/it]
 65%|██████▌   | 820/1261 [3:04:21<1:39:15, 13.50s/it]
 65%|██████▌   | 821/1261 [3:04:35<1:39:09, 13.52s/it]
 65%|██████▌   | 822/1261 [3:04:48<1:38:49, 13.51s/it]
 65%|██████▌   | 823/1261 [3:05:02<1:38:43, 13.53s/it]
 65%|██████▌   | 824/1261 [3:05:15<1:38:25, 13.51s/it]
 65%|██████▌   | 825/1261 [3:05:29<1:38:18, 13.53s/it]
 66%|██████▌   | 826/1261 [3:05:42<1:37:56, 13.51s/it]
 66%|██████▌   | 827/1261 [3:05:56<1:37:49, 13.53s/it]
 66%|██████▌   | 828/1261 [3:06:09<1:37:30, 13.51s/it]
 66%|██████▌   | 829/1261 [3:06:23<1:37:24, 13.53s/it]
 66%|██████▌   | 830/1261 [3:06:37<1:37:03, 13.51s/it]
 66%|██████▌   | 831/1261 [3:06:50<1:36:58, 13.53s/it]
 66%|██████▌   | 832/1261 [3:07:04<1:36:39, 13.52s/it]
 66%|██████▌   | 833/1261 [3:07:17<1:36:32, 13.53s/it]
 66%|██████▌   | 834/1261 [3:07:31<1:36:11, 13.52s/it]
 66%|██████▌   | 835/1261 [3:07:44<1:36:05, 13.53s/it]
 66%|██████▋   | 836/1261 [3:07:58<1:35:45, 13.52s/it]
 66%|██████▋   | 837/1261 [3:08:11<1:35:38, 13.53s/it]
 66%|██████▋   | 838/1261 [3:08:25<1:35:18, 13.52s/it]
 67%|██████▋   | 839/1261 [3:08:38<1:35:12, 13.54s/it]
 67%|██████▋   | 840/1261 [3:08:52<1:34:52, 13.52s/it]
 67%|██████▋   | 841/1261 [3:09:05<1:34:48, 13.54s/it]
 67%|██████▋   | 842/1261 [3:09:19<1:34:31, 13.53s/it]
 67%|██████▋   | 843/1261 [3:09:32<1:34:23, 13.55s/it]
 67%|██████▋   | 844/1261 [3:09:46<1:34:02, 13.53s/it]
 67%|██████▋   | 845/1261 [3:10:00<1:33:53, 13.54s/it]
 67%|██████▋   | 846/1261 [3:10:13<1:33:31, 13.52s/it]
 67%|██████▋   | 847/1261 [3:10:27<1:33:23, 13.53s/it]
 67%|██████▋   | 848/1261 [3:10:40<1:33:03, 13.52s/it]
 67%|██████▋   | 849/1261 [3:10:54<1:32:56, 13.53s/it]
 67%|██████▋   | 850/1261 [3:11:07<1:32:36, 13.52s/it]
 67%|██████▋   | 851/1261 [3:11:21<1:32:30, 13.54s/it]
 68%|██████▊   | 852/1261 [3:11:34<1:32:10, 13.52s/it]
 68%|██████▊   | 853/1261 [3:11:48<1:32:03, 13.54s/it]
 68%|██████▊   | 854/1261 [3:12:01<1:31:43, 13.52s/it]
 68%|██████▊   | 855/1261 [3:12:15<1:31:37, 13.54s/it]
 68%|██████▊   | 856/1261 [3:12:28<1:31:16, 13.52s/it]
 68%|██████▊   | 857/1261 [3:12:42<1:31:09, 13.54s/it]
 68%|██████▊   | 858/1261 [3:12:55<1:30:49, 13.52s/it]
 68%|██████▊   | 859/1261 [3:13:09<1:30:41, 13.54s/it]
 68%|██████▊   | 860/1261 [3:13:22<1:30:21, 13.52s/it]
 68%|██████▊   | 861/1261 [3:13:36<1:30:14, 13.54s/it]
 68%|██████▊   | 862/1261 [3:13:49<1:29:54, 13.52s/it]
 68%|██████▊   | 863/1261 [3:14:03<1:29:47, 13.54s/it]
 69%|██████▊   | 864/1261 [3:14:17<1:29:28, 13.52s/it]
 69%|██████▊   | 865/1261 [3:14:30<1:29:20, 13.54s/it]
 69%|██████▊   | 866/1261 [3:14:44<1:29:00, 13.52s/it]
 69%|██████▉   | 867/1261 [3:14:57<1:28:53, 13.54s/it]
 69%|██████▉   | 868/1261 [3:15:11<1:28:34, 13.52s/it]
 69%|██████▉   | 869/1261 [3:15:24<1:28:26, 13.54s/it]
 69%|██████▉   | 870/1261 [3:15:38<1:28:07, 13.52s/it]
 69%|██████▉   | 871/1261 [3:15:51<1:28:02, 13.55s/it]
 69%|██████▉   | 872/1261 [3:16:05<1:27:41, 13.53s/it]
 69%|██████▉   | 873/1261 [3:16:18<1:27:33, 13.54s/it]
 69%|██████▉   | 874/1261 [3:16:32<1:27:12, 13.52s/it]
 69%|██████▉   | 875/1261 [3:16:45<1:27:04, 13.53s/it]
 69%|██████▉   | 876/1261 [3:16:59<1:26:43, 13.52s/it]
 70%|██████▉   | 877/1261 [3:17:12<1:26:36, 13.53s/it]
 70%|██████▉   | 878/1261 [3:17:26<1:26:16, 13.51s/it]
 70%|██████▉   | 879/1261 [3:17:40<1:26:11, 13.54s/it]
 70%|██████▉   | 880/1261 [3:17:53<1:25:49, 13.52s/it]
 70%|██████▉   | 881/1261 [3:18:07<1:25:41, 13.53s/it]
 70%|██████▉   | 882/1261 [3:18:20<1:25:20, 13.51s/it]
 70%|███████   | 883/1261 [3:18:34<1:25:13, 13.53s/it]
 70%|███████   | 884/1261 [3:18:47<1:24:54, 13.51s/it]
 70%|███████   | 885/1261 [3:19:01<1:24:45, 13.53s/it]
 70%|███████   | 886/1261 [3:19:14<1:24:26, 13.51s/it]
 70%|███████   | 887/1261 [3:19:28<1:24:18, 13.53s/it]
 70%|███████   | 888/1261 [3:19:41<1:23:58, 13.51s/it]
 70%|███████   | 889/1261 [3:19:55<1:23:50, 13.52s/it]
 71%|███████   | 890/1261 [3:20:08<1:23:31, 13.51s/it]
 71%|███████   | 891/1261 [3:20:22<1:23:24, 13.53s/it]
 71%|███████   | 892/1261 [3:20:35<1:23:03, 13.51s/it]
 71%|███████   | 893/1261 [3:20:49<1:22:55, 13.52s/it]
 71%|███████   | 894/1261 [3:21:02<1:22:35, 13.50s/it]
 71%|███████   | 895/1261 [3:21:16<1:22:31, 13.53s/it]
 71%|███████   | 896/1261 [3:21:29<1:22:12, 13.51s/it]
 71%|███████   | 897/1261 [3:21:43<1:22:04, 13.53s/it]
 71%|███████   | 898/1261 [3:21:56<1:21:45, 13.51s/it]
 71%|███████▏  | 899/1261 [3:22:10<1:21:38, 13.53s/it]
 71%|███████▏  | 900/1261 [3:22:23<1:21:19, 13.52s/it]
 71%|███████▏  | 901/1261 [3:22:37<1:21:14, 13.54s/it]
 72%|███████▏  | 902/1261 [3:22:50<1:20:55, 13.53s/it]
 72%|███████▏  | 903/1261 [3:23:04<1:20:46, 13.54s/it]
 72%|███████▏  | 904/1261 [3:23:17<1:20:25, 13.52s/it]
 72%|███████▏  | 905/1261 [3:23:31<1:20:17, 13.53s/it]
 72%|███████▏  | 906/1261 [3:23:44<1:19:56, 13.51s/it]
 72%|███████▏  | 907/1261 [3:23:58<1:19:47, 13.53s/it]
 72%|███████▏  | 908/1261 [3:24:12<1:19:28, 13.51s/it]
 72%|███████▏  | 909/1261 [3:24:25<1:19:19, 13.52s/it]
 72%|███████▏  | 910/1261 [3:24:39<1:19:01, 13.51s/it]
 72%|███████▏  | 911/1261 [3:24:52<1:18:53, 13.53s/it]
 72%|███████▏  | 912/1261 [3:25:06<1:18:33, 13.51s/it]
 72%|███████▏  | 913/1261 [3:25:19<1:18:25, 13.52s/it]
 72%|███████▏  | 914/1261 [3:25:33<1:18:06, 13.51s/it]
 73%|███████▎  | 915/1261 [3:25:46<1:18:02, 13.53s/it]
 73%|███████▎  | 916/1261 [3:26:00<1:17:43, 13.52s/it]
 73%|███████▎  | 917/1261 [3:26:13<1:17:34, 13.53s/it]
 73%|███████▎  | 918/1261 [3:26:27<1:17:14, 13.51s/it]
 73%|███████▎  | 919/1261 [3:26:40<1:17:06, 13.53s/it]
 73%|███████▎  | 920/1261 [3:26:54<1:16:48, 13.51s/it]
 73%|███████▎  | 921/1261 [3:27:07<1:16:39, 13.53s/it]
 73%|███████▎  | 922/1261 [3:27:21<1:16:20, 13.51s/it]
 73%|███████▎  | 923/1261 [3:27:34<1:16:11, 13.53s/it]
 73%|███████▎  | 924/1261 [3:27:48<1:15:52, 13.51s/it]
 73%|███████▎  | 925/1261 [3:28:01<1:15:45, 13.53s/it]
 73%|███████▎  | 926/1261 [3:28:15<1:15:26, 13.51s/it]
 74%|███████▎  | 927/1261 [3:28:28<1:15:18, 13.53s/it]
 74%|███████▎  | 928/1261 [3:28:42<1:15:00, 13.51s/it]
 74%|███████▎  | 929/1261 [3:28:55<1:14:52, 13.53s/it]
 74%|███████▍  | 930/1261 [3:29:09<1:14:34, 13.52s/it]
 74%|███████▍  | 931/1261 [3:29:23<1:14:31, 13.55s/it]
 74%|███████▍  | 932/1261 [3:29:36<1:14:13, 13.54s/it]
 74%|███████▍  | 933/1261 [3:29:50<1:14:04, 13.55s/it]
 74%|███████▍  | 934/1261 [3:30:03<1:13:45, 13.53s/it]
 74%|███████▍  | 935/1261 [3:30:17<1:13:36, 13.55s/it]
 74%|███████▍  | 936/1261 [3:30:30<1:13:15, 13.52s/it]
 74%|███████▍  | 937/1261 [3:30:44<1:13:05, 13.53s/it]
 74%|███████▍  | 938/1261 [3:30:57<1:12:45, 13.52s/it]
 74%|███████▍  | 939/1261 [3:31:11<1:12:37, 13.53s/it]
 75%|███████▍  | 940/1261 [3:31:24<1:12:17, 13.51s/it]
 75%|███████▍  | 941/1261 [3:31:38<1:12:08, 13.53s/it]
 75%|███████▍  | 942/1261 [3:31:51<1:11:48, 13.51s/it]
 75%|███████▍  | 943/1261 [3:32:05<1:11:40, 13.52s/it]
 75%|███████▍  | 944/1261 [3:32:18<1:11:21, 13.51s/it]
 75%|███████▍  | 945/1261 [3:32:32<1:11:13, 13.52s/it]
 75%|███████▌  | 946/1261 [3:32:45<1:10:54, 13.51s/it]
 75%|███████▌  | 947/1261 [3:32:59<1:10:46, 13.52s/it]
 75%|███████▌  | 948/1261 [3:33:12<1:10:27, 13.51s/it]
 75%|███████▌  | 949/1261 [3:33:26<1:10:18, 13.52s/it]
 75%|███████▌  | 950/1261 [3:33:39<1:10:00, 13.51s/it]
 75%|███████▌  | 951/1261 [3:33:53<1:09:51, 13.52s/it]
 75%|███████▌  | 952/1261 [3:34:06<1:09:32, 13.50s/it]
 76%|███████▌  | 953/1261 [3:34:20<1:09:24, 13.52s/it]
 76%|███████▌  | 954/1261 [3:34:33<1:09:06, 13.51s/it]
 76%|███████▌  | 955/1261 [3:34:47<1:08:58, 13.52s/it]
 76%|███████▌  | 956/1261 [3:35:01<1:08:39, 13.51s/it]
 76%|███████▌  | 957/1261 [3:35:14<1:08:32, 13.53s/it]
 76%|███████▌  | 958/1261 [3:35:28<1:08:13, 13.51s/it]
 76%|███████▌  | 959/1261 [3:35:41<1:08:04, 13.52s/it]
 76%|███████▌  | 960/1261 [3:35:55<1:07:45, 13.51s/it]
 76%|███████▌  | 961/1261 [3:36:08<1:07:39, 13.53s/it]
 76%|███████▋  | 962/1261 [3:36:22<1:07:21, 13.52s/it]
 76%|███████▋  | 963/1261 [3:36:35<1:07:14, 13.54s/it]
 76%|███████▋  | 964/1261 [3:36:49<1:06:56, 13.52s/it]
 77%|███████▋  | 965/1261 [3:37:02<1:06:45, 13.53s/it]
 77%|███████▋  | 966/1261 [3:37:16<1:06:26, 13.51s/it]
 77%|███████▋  | 967/1261 [3:37:29<1:06:17, 13.53s/it]
 77%|███████▋  | 968/1261 [3:37:43<1:05:58, 13.51s/it]
 77%|███████▋  | 969/1261 [3:37:56<1:05:49, 13.53s/it]
 77%|███████▋  | 970/1261 [3:38:10<1:05:32, 13.51s/it]
 77%|███████▋  | 971/1261 [3:38:23<1:05:23, 13.53s/it]
 77%|███████▋  | 972/1261 [3:38:37<1:05:05, 13.51s/it]
 77%|███████▋  | 973/1261 [3:38:50<1:04:57, 13.53s/it]
 77%|███████▋  | 974/1261 [3:39:04<1:04:38, 13.51s/it]
 77%|███████▋  | 975/1261 [3:39:18<1:04:30, 13.53s/it]
 77%|███████▋  | 976/1261 [3:39:31<1:04:12, 13.52s/it]
 77%|███████▋  | 977/1261 [3:39:45<1:04:03, 13.53s/it]
 78%|███████▊  | 978/1261 [3:39:58<1:03:46, 13.52s/it]
 78%|███████▊  | 979/1261 [3:40:12<1:03:38, 13.54s/it]
 78%|███████▊  | 980/1261 [3:40:25<1:03:22, 13.53s/it]
 78%|███████▊  | 981/1261 [3:40:39<1:03:11, 13.54s/it]
 78%|███████▊  | 982/1261 [3:40:52<1:02:53, 13.53s/it]
 78%|███████▊  | 983/1261 [3:41:06<1:02:44, 13.54s/it]
 78%|███████▊  | 984/1261 [3:41:19<1:02:26, 13.52s/it]
 78%|███████▊  | 985/1261 [3:41:33<1:02:16, 13.54s/it]
 78%|███████▊  | 986/1261 [3:41:46<1:01:58, 13.52s/it]
 78%|███████▊  | 987/1261 [3:42:00<1:01:48, 13.54s/it]
 78%|███████▊  | 988/1261 [3:42:13<1:01:29, 13.52s/it]
 78%|███████▊  | 989/1261 [3:42:27<1:01:20, 13.53s/it]
 79%|███████▊  | 990/1261 [3:42:40<1:01:02, 13.51s/it]
 79%|███████▊  | 991/1261 [3:42:54<1:00:53, 13.53s/it]
 79%|███████▊  | 992/1261 [3:43:07<1:00:36, 13.52s/it]
 79%|███████▊  | 993/1261 [3:43:21<1:00:27, 13.53s/it]
 79%|███████▉  | 994/1261 [3:43:35<1:00:10, 13.52s/it]
 79%|███████▉  | 995/1261 [3:43:48<1:00:01, 13.54s/it]
 79%|███████▉  | 996/1261 [3:44:02<59:43, 13.52s/it]  
 79%|███████▉  | 997/1261 [3:44:15<59:34, 13.54s/it]
 79%|███████▉  | 998/1261 [3:44:29<59:17, 13.53s/it]
 79%|███████▉  | 999/1261 [3:44:42<59:08, 13.54s/it]
 79%|███████▉  | 1000/1261 [3:44:56<58:50, 13.53s/it]
 79%|███████▉  | 1001/1261 [3:45:09<58:42, 13.55s/it]
 79%|███████▉  | 1002/1261 [3:45:23<58:24, 13.53s/it]
 80%|███████▉  | 1003/1261 [3:45:36<58:14, 13.55s/it]
 80%|███████▉  | 1004/1261 [3:45:50<57:56, 13.53s/it]
 80%|███████▉  | 1005/1261 [3:46:03<57:46, 13.54s/it]
 80%|███████▉  | 1006/1261 [3:46:17<57:28, 13.53s/it]
 80%|███████▉  | 1007/1261 [3:46:31<57:19, 13.54s/it]
 80%|███████▉  | 1008/1261 [3:46:44<57:02, 13.53s/it]
 80%|████████  | 1009/1261 [3:46:58<56:52, 13.54s/it]
 80%|████████  | 1010/1261 [3:47:11<56:35, 13.53s/it]
 80%|████████  | 1011/1261 [3:47:25<56:26, 13.55s/it]
 80%|████████  | 1012/1261 [3:47:38<56:08, 13.53s/it]
 80%|████████  | 1013/1261 [3:47:52<55:59, 13.55s/it]
 80%|████████  | 1014/1261 [3:48:05<55:41, 13.53s/it]
 80%|████████  | 1015/1261 [3:48:19<55:31, 13.54s/it]
 81%|████████  | 1016/1261 [3:48:32<55:14, 13.53s/it]
 81%|████████  | 1017/1261 [3:48:46<55:04, 13.54s/it]
 81%|████████  | 1018/1261 [3:48:59<54:46, 13.53s/it]
 81%|████████  | 1019/1261 [3:49:13<54:37, 13.54s/it]
 81%|████████  | 1020/1261 [3:49:26<54:19, 13.52s/it]
 81%|████████  | 1021/1261 [3:49:40<54:08, 13.54s/it]
 81%|████████  | 1022/1261 [3:49:53<53:51, 13.52s/it]
 81%|████████  | 1023/1261 [3:50:07<53:40, 13.53s/it]
 81%|████████  | 1024/1261 [3:50:21<53:23, 13.52s/it]
 81%|████████▏ | 1025/1261 [3:50:34<53:13, 13.53s/it]
 81%|████████▏ | 1026/1261 [3:50:48<52:56, 13.52s/it]
 81%|████████▏ | 1027/1261 [3:51:01<52:45, 13.53s/it]
 82%|████████▏ | 1028/1261 [3:51:15<52:28, 13.51s/it]
 82%|████████▏ | 1029/1261 [3:51:28<52:18, 13.53s/it]
 82%|████████▏ | 1030/1261 [3:51:42<52:02, 13.52s/it]
 82%|████████▏ | 1031/1261 [3:51:55<51:52, 13.53s/it]
 82%|████████▏ | 1032/1261 [3:52:09<51:34, 13.51s/it]
 82%|████████▏ | 1033/1261 [3:52:22<51:23, 13.53s/it]
 82%|████████▏ | 1034/1261 [3:52:36<51:05, 13.51s/it]
 82%|████████▏ | 1035/1261 [3:52:49<50:55, 13.52s/it]
 82%|████████▏ | 1036/1261 [3:53:03<50:37, 13.50s/it]
 82%|████████▏ | 1037/1261 [3:53:16<50:27, 13.52s/it]
 82%|████████▏ | 1038/1261 [3:53:30<50:10, 13.50s/it]
 82%|████████▏ | 1039/1261 [3:53:43<50:00, 13.52s/it]
 82%|████████▏ | 1040/1261 [3:53:57<49:43, 13.50s/it]
 83%|████████▎ | 1041/1261 [3:54:10<49:33, 13.52s/it]
 83%|████████▎ | 1042/1261 [3:54:24<49:16, 13.50s/it]
 83%|████████▎ | 1043/1261 [3:54:37<49:06, 13.52s/it]
 83%|████████▎ | 1044/1261 [3:54:51<48:50, 13.50s/it]
 83%|████████▎ | 1045/1261 [3:55:04<48:40, 13.52s/it]
 83%|████████▎ | 1046/1261 [3:55:18<48:23, 13.50s/it]
 83%|████████▎ | 1047/1261 [3:55:31<48:13, 13.52s/it]
 83%|████████▎ | 1048/1261 [3:55:45<47:56, 13.50s/it]
 83%|████████▎ | 1049/1261 [3:55:58<47:46, 13.52s/it]
 83%|████████▎ | 1050/1261 [3:56:12<47:30, 13.51s/it]
 83%|████████▎ | 1051/1261 [3:56:25<47:19, 13.52s/it]
 83%|████████▎ | 1052/1261 [3:56:39<47:03, 13.51s/it]
 84%|████████▎ | 1053/1261 [3:56:53<46:54, 13.53s/it]
 84%|████████▎ | 1054/1261 [3:57:06<46:36, 13.51s/it]
 84%|████████▎ | 1055/1261 [3:57:20<46:27, 13.53s/it]
 84%|████████▎ | 1056/1261 [3:57:33<46:10, 13.51s/it]
 84%|████████▍ | 1057/1261 [3:57:47<46:01, 13.54s/it]
 84%|████████▍ | 1058/1261 [3:58:00<45:45, 13.52s/it]
 84%|████████▍ | 1059/1261 [3:58:14<45:34, 13.54s/it]
 84%|████████▍ | 1060/1261 [3:58:27<45:18, 13.53s/it]
 84%|████████▍ | 1061/1261 [3:58:41<45:08, 13.54s/it]
 84%|████████▍ | 1062/1261 [3:58:54<44:52, 13.53s/it]
 84%|████████▍ | 1063/1261 [3:59:08<44:42, 13.55s/it]
 84%|████████▍ | 1064/1261 [3:59:21<44:26, 13.53s/it]
 84%|████████▍ | 1065/1261 [3:59:35<44:16, 13.55s/it]
 85%|████████▍ | 1066/1261 [3:59:48<43:59, 13.53s/it]
 85%|████████▍ | 1067/1261 [4:00:02<43:48, 13.55s/it]
 85%|████████▍ | 1068/1261 [4:00:16<43:32, 13.53s/it]
 85%|████████▍ | 1069/1261 [4:00:29<43:21, 13.55s/it]
 85%|████████▍ | 1070/1261 [4:00:43<43:04, 13.53s/it]
 85%|████████▍ | 1071/1261 [4:00:56<42:54, 13.55s/it]
 85%|████████▌ | 1072/1261 [4:01:10<42:38, 13.54s/it]
 85%|████████▌ | 1073/1261 [4:01:23<42:28, 13.56s/it]
 85%|████████▌ | 1074/1261 [4:01:37<42:13, 13.55s/it]
 85%|████████▌ | 1075/1261 [4:01:50<42:02, 13.56s/it]
 85%|████████▌ | 1076/1261 [4:02:04<41:46, 13.55s/it]
 85%|████████▌ | 1077/1261 [4:02:18<41:35, 13.56s/it]
 85%|████████▌ | 1078/1261 [4:02:31<41:18, 13.55s/it]
 86%|████████▌ | 1079/1261 [4:02:45<41:08, 13.56s/it]
 86%|████████▌ | 1080/1261 [4:02:58<40:51, 13.54s/it]
 86%|████████▌ | 1081/1261 [4:03:12<40:39, 13.55s/it]
 86%|████████▌ | 1082/1261 [4:03:25<40:22, 13.53s/it]
 86%|████████▌ | 1083/1261 [4:03:39<40:10, 13.54s/it]
 86%|████████▌ | 1084/1261 [4:03:52<39:54, 13.53s/it]
 86%|████████▌ | 1085/1261 [4:04:06<39:44, 13.55s/it]
 86%|████████▌ | 1086/1261 [4:04:19<39:28, 13.53s/it]
 86%|████████▌ | 1087/1261 [4:04:33<39:18, 13.55s/it]
 86%|████████▋ | 1088/1261 [4:04:46<39:02, 13.54s/it]
 86%|████████▋ | 1089/1261 [4:05:00<38:51, 13.56s/it]
 86%|████████▋ | 1090/1261 [4:05:14<38:35, 13.54s/it]
 87%|████████▋ | 1091/1261 [4:05:27<38:24, 13.56s/it]
 87%|████████▋ | 1092/1261 [4:05:41<38:09, 13.55s/it]
 87%|████████▋ | 1093/1261 [4:05:54<37:58, 13.56s/it]
 87%|████████▋ | 1094/1261 [4:06:08<37:41, 13.54s/it]
 87%|████████▋ | 1095/1261 [4:06:21<37:30, 13.56s/it]
 87%|████████▋ | 1096/1261 [4:06:35<37:14, 13.54s/it]
 87%|████████▋ | 1097/1261 [4:06:48<37:03, 13.56s/it]
 87%|████████▋ | 1098/1261 [4:07:02<36:47, 13.54s/it]
 87%|████████▋ | 1099/1261 [4:07:16<36:36, 13.56s/it]
 87%|████████▋ | 1100/1261 [4:07:29<36:20, 13.54s/it]
 87%|████████▋ | 1101/1261 [4:07:43<36:08, 13.56s/it]
 87%|████████▋ | 1102/1261 [4:07:56<35:52, 13.54s/it]
 87%|████████▋ | 1103/1261 [4:08:10<35:41, 13.56s/it]
 88%|████████▊ | 1104/1261 [4:08:23<35:25, 13.54s/it]
 88%|████████▊ | 1105/1261 [4:08:37<35:15, 13.56s/it]
 88%|████████▊ | 1106/1261 [4:08:50<34:58, 13.54s/it]
 88%|████████▊ | 1107/1261 [4:09:04<34:47, 13.55s/it]
 88%|████████▊ | 1108/1261 [4:09:17<34:31, 13.54s/it]
 88%|████████▊ | 1109/1261 [4:09:31<34:19, 13.55s/it]
 88%|████████▊ | 1110/1261 [4:09:45<34:03, 13.53s/it]
 88%|████████▊ | 1111/1261 [4:09:58<33:52, 13.55s/it]
 88%|████████▊ | 1112/1261 [4:10:12<33:36, 13.54s/it]
 88%|████████▊ | 1113/1261 [4:10:25<33:25, 13.55s/it]
 88%|████████▊ | 1114/1261 [4:10:39<33:09, 13.53s/it]
 88%|████████▊ | 1115/1261 [4:10:52<32:58, 13.55s/it]
 89%|████████▊ | 1116/1261 [4:11:06<32:43, 13.54s/it]
 89%|████████▊ | 1117/1261 [4:11:19<32:32, 13.56s/it]
 89%|████████▊ | 1118/1261 [4:11:33<32:15, 13.54s/it]
 89%|████████▊ | 1119/1261 [4:11:46<32:04, 13.55s/it]
 89%|████████▉ | 1120/1261 [4:12:00<31:48, 13.53s/it]
 89%|████████▉ | 1121/1261 [4:12:14<31:37, 13.55s/it]
 89%|████████▉ | 1122/1261 [4:12:27<31:20, 13.53s/it]
 89%|████████▉ | 1123/1261 [4:12:41<31:09, 13.55s/it]
 89%|████████▉ | 1124/1261 [4:12:54<30:53, 13.53s/it]
 89%|████████▉ | 1125/1261 [4:13:08<30:42, 13.55s/it]
 89%|████████▉ | 1126/1261 [4:13:21<30:26, 13.53s/it]
 89%|████████▉ | 1127/1261 [4:13:35<30:14, 13.54s/it]
 89%|████████▉ | 1128/1261 [4:13:48<29:59, 13.53s/it]
 90%|████████▉ | 1129/1261 [4:14:02<29:47, 13.54s/it]
 90%|████████▉ | 1130/1261 [4:14:15<29:32, 13.53s/it]
 90%|████████▉ | 1131/1261 [4:14:29<29:20, 13.55s/it]
 90%|████████▉ | 1132/1261 [4:14:42<29:05, 13.53s/it]
 90%|████████▉ | 1133/1261 [4:14:56<28:53, 13.54s/it]
 90%|████████▉ | 1134/1261 [4:15:09<28:38, 13.53s/it]
 90%|█████████ | 1135/1261 [4:15:23<28:26, 13.54s/it]
 90%|█████████ | 1136/1261 [4:15:37<28:11, 13.53s/it]
 90%|█████████ | 1137/1261 [4:15:50<27:59, 13.54s/it]
 90%|█████████ | 1138/1261 [4:16:04<27:43, 13.53s/it]
 90%|█████████ | 1139/1261 [4:16:17<27:32, 13.54s/it]
 90%|█████████ | 1140/1261 [4:16:31<27:16, 13.52s/it]
 90%|█████████ | 1141/1261 [4:16:44<27:05, 13.55s/it]
 91%|█████████ | 1142/1261 [4:16:58<26:50, 13.54s/it]
 91%|█████████ | 1143/1261 [4:17:11<26:38, 13.55s/it]
 91%|█████████ | 1144/1261 [4:17:25<26:23, 13.53s/it]
 91%|█████████ | 1145/1261 [4:17:38<26:11, 13.54s/it]
 91%|█████████ | 1146/1261 [4:17:52<25:55, 13.53s/it]
 91%|█████████ | 1147/1261 [4:18:06<25:43, 13.54s/it]
 91%|█████████ | 1148/1261 [4:18:19<25:28, 13.52s/it]
 91%|█████████ | 1149/1261 [4:18:33<25:16, 13.54s/it]
 91%|█████████ | 1150/1261 [4:18:46<25:00, 13.52s/it]
 91%|█████████▏| 1151/1261 [4:19:00<24:49, 13.54s/it]
 91%|█████████▏| 1152/1261 [4:19:13<24:34, 13.52s/it]
 91%|█████████▏| 1153/1261 [4:19:27<24:22, 13.54s/it]
 92%|█████████▏| 1154/1261 [4:19:40<24:07, 13.53s/it]
 92%|█████████▏| 1155/1261 [4:19:54<23:55, 13.54s/it]
 92%|█████████▏| 1156/1261 [4:20:07<23:39, 13.52s/it]
 92%|█████████▏| 1157/1261 [4:20:21<23:27, 13.53s/it]
 92%|█████████▏| 1158/1261 [4:20:34<23:12, 13.52s/it]
 92%|█████████▏| 1159/1261 [4:20:48<23:00, 13.54s/it]
 92%|█████████▏| 1160/1261 [4:21:01<22:45, 13.52s/it]
 92%|█████████▏| 1161/1261 [4:21:15<22:33, 13.53s/it]
 92%|█████████▏| 1162/1261 [4:21:28<22:18, 13.52s/it]
 92%|█████████▏| 1163/1261 [4:21:42<22:06, 13.53s/it]
 92%|█████████▏| 1164/1261 [4:21:55<21:51, 13.52s/it]
 92%|█████████▏| 1165/1261 [4:22:09<21:39, 13.53s/it]
 92%|█████████▏| 1166/1261 [4:22:23<21:24, 13.52s/it]
 93%|█████████▎| 1167/1261 [4:22:36<21:12, 13.54s/it]
 93%|█████████▎| 1168/1261 [4:22:50<20:57, 13.52s/it]
 93%|█████████▎| 1169/1261 [4:23:03<20:45, 13.54s/it]
 93%|█████████▎| 1170/1261 [4:23:17<20:30, 13.53s/it]
 93%|█████████▎| 1171/1261 [4:23:30<20:19, 13.55s/it]
 93%|█████████▎| 1172/1261 [4:23:44<20:04, 13.53s/it]
 93%|█████████▎| 1173/1261 [4:23:57<19:52, 13.55s/it]
 93%|█████████▎| 1174/1261 [4:24:11<19:37, 13.53s/it]
 93%|█████████▎| 1175/1261 [4:24:24<19:24, 13.54s/it]
 93%|█████████▎| 1176/1261 [4:24:38<19:09, 13.53s/it]
 93%|█████████▎| 1177/1261 [4:24:51<18:57, 13.54s/it]
 93%|█████████▎| 1178/1261 [4:25:05<18:42, 13.52s/it]
 93%|█████████▎| 1179/1261 [4:25:18<18:29, 13.53s/it]
 94%|█████████▎| 1180/1261 [4:25:32<18:14, 13.52s/it]
 94%|█████████▎| 1181/1261 [4:25:46<18:03, 13.54s/it]
 94%|█████████▎| 1182/1261 [4:25:59<17:48, 13.52s/it]
 94%|█████████▍| 1183/1261 [4:26:13<17:35, 13.53s/it]
 94%|█████████▍| 1184/1261 [4:26:26<17:20, 13.51s/it]
 94%|█████████▍| 1185/1261 [4:26:40<17:07, 13.53s/it]
 94%|█████████▍| 1186/1261 [4:26:53<16:53, 13.51s/it]
 94%|█████████▍| 1187/1261 [4:27:07<16:40, 13.52s/it]
 94%|█████████▍| 1188/1261 [4:27:20<16:26, 13.51s/it]
 94%|█████████▍| 1189/1261 [4:27:34<16:13, 13.52s/it]
 94%|█████████▍| 1190/1261 [4:27:47<15:58, 13.51s/it]
 94%|█████████▍| 1191/1261 [4:28:01<15:46, 13.52s/it]
 95%|█████████▍| 1192/1261 [4:28:14<15:31, 13.51s/it]
 95%|█████████▍| 1193/1261 [4:28:28<15:19, 13.52s/it]
 95%|█████████▍| 1194/1261 [4:28:41<15:04, 13.51s/it]
 95%|█████████▍| 1195/1261 [4:28:55<14:52, 13.52s/it]
 95%|█████████▍| 1196/1261 [4:29:08<14:37, 13.51s/it]
 95%|█████████▍| 1197/1261 [4:29:22<14:25, 13.53s/it]
 95%|█████████▌| 1198/1261 [4:29:35<14:11, 13.51s/it]
 95%|█████████▌| 1199/1261 [4:29:49<13:58, 13.53s/it]
 95%|█████████▌| 1200/1261 [4:30:02<13:43, 13.51s/it]
 95%|█████████▌| 1201/1261 [4:30:16<13:31, 13.53s/it]
 95%|█████████▌| 1202/1261 [4:30:29<13:17, 13.51s/it]
 95%|█████████▌| 1203/1261 [4:30:43<13:04, 13.53s/it]
 95%|█████████▌| 1204/1261 [4:30:56<12:50, 13.51s/it]
 96%|█████████▌| 1205/1261 [4:31:10<12:37, 13.53s/it]
 96%|█████████▌| 1206/1261 [4:31:23<12:23, 13.51s/it]
 96%|█████████▌| 1207/1261 [4:31:37<12:10, 13.53s/it]
 96%|█████████▌| 1208/1261 [4:31:50<11:56, 13.51s/it]
 96%|█████████▌| 1209/1261 [4:32:04<11:43, 13.53s/it]
 96%|█████████▌| 1210/1261 [4:32:18<11:29, 13.51s/it]
 96%|█████████▌| 1211/1261 [4:32:31<11:16, 13.53s/it]
 96%|█████████▌| 1212/1261 [4:32:45<11:02, 13.52s/it]
 96%|█████████▌| 1213/1261 [4:32:58<10:49, 13.53s/it]
 96%|█████████▋| 1214/1261 [4:33:12<10:35, 13.51s/it]
 96%|█████████▋| 1215/1261 [4:33:25<10:22, 13.53s/it]
 96%|█████████▋| 1216/1261 [4:33:39<10:08, 13.51s/it]
 97%|█████████▋| 1217/1261 [4:33:52<09:55, 13.53s/it]
 97%|█████████▋| 1218/1261 [4:34:06<09:40, 13.51s/it]
 97%|█████████▋| 1219/1261 [4:34:19<09:28, 13.53s/it]
 97%|█████████▋| 1220/1261 [4:34:33<09:14, 13.51s/it]
 97%|█████████▋| 1221/1261 [4:34:46<09:01, 13.53s/it]
 97%|█████████▋| 1222/1261 [4:35:00<08:47, 13.51s/it]
 97%|█████████▋| 1223/1261 [4:35:13<08:34, 13.53s/it]
 97%|█████████▋| 1224/1261 [4:35:27<08:19, 13.51s/it]
 97%|█████████▋| 1225/1261 [4:35:40<08:06, 13.53s/it]
 97%|█████████▋| 1226/1261 [4:35:54<07:52, 13.51s/it]
 97%|█████████▋| 1227/1261 [4:36:07<07:39, 13.53s/it]
 97%|█████████▋| 1228/1261 [4:36:21<07:25, 13.51s/it]
 97%|█████████▋| 1229/1261 [4:36:34<07:12, 13.53s/it]
 98%|█████████▊| 1230/1261 [4:36:48<06:59, 13.52s/it]
 98%|█████████▊| 1231/1261 [4:37:02<06:46, 13.54s/it]
 98%|█████████▊| 1232/1261 [4:37:15<06:32, 13.52s/it]
 98%|█████████▊| 1233/1261 [4:37:29<06:19, 13.54s/it]
 98%|█████████▊| 1234/1261 [4:37:42<06:05, 13.52s/it]
 98%|█████████▊| 1235/1261 [4:37:56<05:51, 13.54s/it]
 98%|█████████▊| 1236/1261 [4:38:09<05:38, 13.52s/it]
 98%|█████████▊| 1237/1261 [4:38:23<05:24, 13.54s/it]
 98%|█████████▊| 1238/1261 [4:38:36<05:10, 13.52s/it]
 98%|█████████▊| 1239/1261 [4:38:50<04:57, 13.54s/it]
 98%|█████████▊| 1240/1261 [4:39:03<04:43, 13.52s/it]
 98%|█████████▊| 1241/1261 [4:39:17<04:30, 13.53s/it]
 98%|█████████▊| 1242/1261 [4:39:30<04:16, 13.52s/it]
 99%|█████████▊| 1243/1261 [4:39:44<04:03, 13.53s/it]
 99%|█████████▊| 1244/1261 [4:39:57<03:49, 13.52s/it]
 99%|█████████▊| 1245/1261 [4:40:11<03:36, 13.53s/it]
 99%|█████████▉| 1246/1261 [4:40:24<03:22, 13.51s/it]
 99%|█████████▉| 1247/1261 [4:40:38<03:09, 13.53s/it]
 99%|█████████▉| 1248/1261 [4:40:51<02:55, 13.51s/it]
 99%|█████████▉| 1249/1261 [4:41:05<02:42, 13.53s/it]
 99%|█████████▉| 1250/1261 [4:41:18<02:28, 13.51s/it]
 99%|█████████▉| 1251/1261 [4:41:32<02:15, 13.53s/it]
 99%|█████████▉| 1252/1261 [4:41:46<02:01, 13.51s/it]
 99%|█████████▉| 1253/1261 [4:41:59<01:48, 13.54s/it]
 99%|█████████▉| 1254/1261 [4:42:13<01:34, 13.52s/it]
100%|█████████▉| 1255/1261 [4:42:26<01:21, 13.54s/it]
100%|█████████▉| 1256/1261 [4:42:40<01:07, 13.52s/it]
100%|█████████▉| 1257/1261 [4:42:53<00:54, 13.54s/it]
100%|█████████▉| 1258/1261 [4:43:07<00:40, 13.52s/it]
100%|█████████▉| 1259/1261 [4:43:20<00:27, 13.54s/it]
100%|█████████▉| 1260/1261 [4:43:34<00:13, 13.52s/it]

[MoviePy] Done.
[MoviePy] >>>> Video ready: project_video_result.mp4 

CPU times: user 4h 43min 36s, sys: 1.45 s, total: 4h 43min 37s
Wall time: 4h 43min 34s
In [58]:
HTML("""
<video width="960" height="540" controls>
  <source src="{0}">
</video>
""".format(output_clip))
Out[58]:

Challenges faced and future exploration

Removing False positives was the biggest challenge. There are still some False positives in the final video and pipeline needs to be further improved. Weighted moving average could be used to further reduce the flickering in the video.

As next steps I would like to explore Facebook Research's MultiPathNet Object Detection Deep Learning model for vehicle detection.

The pipeline is also very time consuming, taking many hours to generate the final video even when running on aws p2.8xlarge machines. Techiques for optimizing and parallelizing the pipeline are needed especially for the vehicle detection part, otherwise things will not work in real-time for a real autonomous car. Reducing the search area for the search window and Facebook's MultiPathNet could be possible options for optimizing.

In [ ]: